Document Promptkit v0.5.0

README.md

@@ -33,6 +33,15 @@ boundary and constraints that framework work must preserve.
 
 ## Release Guidance
 
+Consumers upgrading from `v0.4.0` to `v0.5.0` should read the
+[v0.5.0 changelog and migration guide](docs/releases/v0.5.0.md).
+
+Consumers upgrading from `v0.3.0` to `v0.4.0` should read the
+[v0.4.0 changelog and adoption guide](docs/releases/v0.4.0.md).
+
+Consumers upgrading from `v0.2.0` to `v0.3.0` should read the
+[v0.3.0 changelog](docs/releases/v0.3.0.md).
+
 Consumers moving from `v0.1.0` to `v0.2.0` should read the
 [v0.2.0 changelog and migration guide](docs/releases/v0.2.0.md).
 

backends.go

@@ -9,6 +9,11 @@ import (
 // backend.
 const BackendOpenRouter = backend.OpenRouterID
 
+// BackendLocal is the case-sensitive conventional ID used by [LocalBackend].
+// It is not a built-in or reserved backend and must be registered with
+// [WithBackend].
+const BackendLocal = "local"
+
 // Backend configures one engine-scoped OpenAI-compatible backend.
 //
 // Backend has no stable JSON representation. Use keyed literals so additions
@@ -35,15 +40,35 @@ type Backend struct {
 	// calls allowed for this backend within one Engine. Zero leaves the backend
 	// unlimited. A negative value makes NewEngine fail with ErrInvalidConfig.
 	ConcurrencyLimit int
-	// QueueCapacity controls how many additional Run calls may be admitted
-	// beyond ConcurrencyLimit. Nil uses 1024 when ConcurrencyLimit is positive;
-	// a pointer uses its exact value, including zero. The pointed-to value must
-	// be non-negative, and QueueCapacity must be nil when ConcurrencyLimit is
-	// zero. Their sum must fit in an int. WithBackend copies the value and does
-	// not retain the pointer.
+	// QueueCapacity controls how many additional Run or RunPrepared calls may
+	// be admitted beyond ConcurrencyLimit. Nil uses 1024 when ConcurrencyLimit
+	// is positive; a pointer uses its exact value, including zero. The pointed-to
+	// value must be non-negative, and QueueCapacity must be nil when
+	// ConcurrencyLimit is zero. Their sum must fit in an int. WithBackend copies
+	// the value and does not retain the pointer.
 	QueueCapacity *int
 }
 
+// LocalBackend returns a caller-owned Backend for a conventional local
+// OpenAI-compatible endpoint. It sets ID to BackendLocal and copies endpoint
+// and concurrencyLimit into Endpoint and ConcurrencyLimit without
+// normalization or validation. APIKeyEnv, ExtraParams, and QueueCapacity keep
+// their zero values.
+//
+// LocalBackend does not read environment variables, register the value, or
+// mutate engine or package state. Supply the returned value through
+// [WithBackend]; [NewEngine] then applies the ordinary backend validation and
+// concurrency semantics, including default queue capacity for a positive
+// limit, unlimited behavior for zero, and ErrInvalidConfig for a negative
+// limit.
+func LocalBackend(endpoint string, concurrencyLimit int) Backend {
+	return Backend{
+		ID:               BackendLocal,
+		Endpoint:         endpoint,
+		ConcurrencyLimit: concurrencyLimit,
+	}
+}
+
 // WithBackend adds one Backend registration to the constructed Engine.
 //
 // Registrations accumulate in option order. Every normalized ID must be unique

capacity_contract_test.go

@@ -60,7 +60,18 @@ func TestEngineRejectsRunBeforeCompletionWhenAdmissionIsFull(t *testing.T) {
 
 	awaitCapacitySignal(t, reader.entered, "first artifact read")
 
-	result, err := engine.Run(context.Background(), capacityInputRequest("http://second.example/v1"))
+	canceledContext, cancel := context.WithCancel(context.Background())
+	cancel()
+	result, err := engine.Run(canceledContext, capacityInputRequest("http://canceled.example/v1"))
+	if result != nil || !errors.Is(err, context.Canceled) {
+		t.Fatalf("canceled capacity admission=(%+v, %v), want context cancellation", result, err)
+	}
+	var canceledCapacityErr *promptkit.CapacityError
+	if errors.Is(err, promptkit.ErrCapacityExceeded) || errors.As(err, &canceledCapacityErr) {
+		t.Fatalf("canceled admission exposed capacity rejection: %v", err)
+	}
+
+	result, err = engine.Run(context.Background(), capacityInputRequest("http://second.example/v1"))
 	if result != nil {
 		t.Fatalf("capacity rejection returned partial result: %+v", result)
 	}
@@ -70,6 +81,21 @@ func TestEngineRejectsRunBeforeCompletionWhenAdmissionIsFull(t *testing.T) {
 	if errors.Is(err, promptkit.ErrInvalidRequest) || errors.Is(err, promptkit.ErrLLMGenerate) {
 		t.Fatalf("capacity rejection had an unrelated category: %v", err)
 	}
+	var capacityErr *promptkit.CapacityError
+	if !errors.As(err, &capacityErr) || capacityErr == nil {
+		t.Fatalf("capacity rejection=%v, want CapacityError", err)
+	}
+	if capacityErr.BackendID != "limited" {
+		t.Fatalf("capacity backend ID=%q, want limited", capacityErr.BackendID)
+	}
+	capacityErr.BackendID = "changed"
+
+	result, err = engine.Run(context.Background(), capacityInputRequest("http://third.example/v1"))
+	var subsequentCapacityErr *promptkit.CapacityError
+	if result != nil || !errors.As(err, &subsequentCapacityErr) ||
+		subsequentCapacityErr == nil || subsequentCapacityErr.BackendID != "limited" {
+		t.Fatalf("subsequent capacity rejection=(%+v, %v), want independent limited CapacityError", result, err)
+	}
 	if calls := reader.callCount(); calls != 1 {
 		t.Fatalf("artifact calls=%d, want only the admitted run", calls)
 	}
@@ -174,6 +200,19 @@ func TestCapacityExceededSentinelContract(t *testing.T) {
 	if promptkit.ErrCapacityExceeded == nil {
 		t.Fatal("ErrCapacityExceeded is nil")
 	}
+	var nilCapacityErr *promptkit.CapacityError
+	zeroCapacityErr := &promptkit.CapacityError{}
+	populatedCapacityErr := &promptkit.CapacityError{BackendID: "limited"}
+	for _, capacityErr := range []error{nilCapacityErr, zeroCapacityErr} {
+		if !errors.Is(capacityErr, promptkit.ErrCapacityExceeded) {
+			t.Fatalf("capacity error=%v, want ErrCapacityExceeded", capacityErr)
+		}
+	}
+	var discoveredCapacityErr *promptkit.CapacityError
+	if !errors.As(populatedCapacityErr, &discoveredCapacityErr) || discoveredCapacityErr != populatedCapacityErr {
+		t.Fatalf("populated capacity error is not discoverable: %v", populatedCapacityErr)
+	}
+
 	for _, unrelated := range []error{
 		promptkit.ErrInvalidConfig,
 		promptkit.ErrInvalidRequest,
@@ -181,7 +220,8 @@ func TestCapacityExceededSentinelContract(t *testing.T) {
 		promptkit.ErrValidation,
 	} {
 		if errors.Is(promptkit.ErrCapacityExceeded, unrelated) ||
-			errors.Is(unrelated, promptkit.ErrCapacityExceeded) {
+			errors.Is(unrelated, promptkit.ErrCapacityExceeded) ||
+			errors.Is(populatedCapacityErr, unrelated) {
 			t.Fatalf("ErrCapacityExceeded aliases unrelated sentinel %v", unrelated)
 		}
 	}

capacity_error.go

@@ -0,0 +1,40 @@
+package promptkit
+
+import (
+	"fmt"
+	"strings"
+)
+
+// CapacityError reports bounded admission rejected for a selected backend.
+//
+// Engine-produced values identify only rejection at Promptkit's bounded
+// [Engine.Run] or [Engine.RunPrepared] admission boundary. BackendID is the
+// normalized registered backend ID used for routing and capacity; endpoint
+// overrides do not change it. Every engine-produced value is nonnil and has a
+// nonblank BackendID. Provider errors, active-generation waiting, and caller
+// cancellation are not represented by this type.
+//
+// Callers own returned values and may mutate BackendID without affecting engine
+// state or another error. CapacityError and its default Go encoding have no
+// stable JSON contract. Consumer-constructed values do not establish that an
+// engine rejected work.
+type CapacityError struct {
+	// BackendID is the normalized registered backend ID whose admission was
+	// rejected.
+	BackendID string
+}
+
+// Error returns diagnostic wording that is not a parsing contract. It is safe
+// to call on a nil receiver or a value with a blank BackendID.
+func (e *CapacityError) Error() string {
+	if e == nil || strings.TrimSpace(e.BackendID) == "" {
+		return ErrCapacityExceeded.Error()
+	}
+	return fmt.Sprintf("backend %q admission: %v", e.BackendID, ErrCapacityExceeded)
+}
+
+// Unwrap returns ErrCapacityExceeded so errors.Is and errors.As can be used
+// together. It is safe to call on a nil receiver or a zero value.
+func (e *CapacityError) Unwrap() error {
+	return ErrCapacityExceeded
+}

convert.go

@@ -170,6 +170,40 @@ func fromDomainExecutionTarget(target domain.ExecutionTarget) ExecutionTarget {
 	}
 }
 
+func fromDomainProfileInspection(inspection *domain.ProfileInspection) *ProfileInspection {
+	if inspection == nil {
+		return nil
+	}
+	return &ProfileInspection{
+		ProfileID:            inspection.ProfileID,
+		EffectiveModelParams: fromDomainExecutionTarget(inspection.EffectiveModelParams),
+		APIKeyRequired:       inspection.APIKeyRequired,
+	}
+}
+
+func fromDomainPromptInspection(inspection *domain.PromptInspection) *PromptInspection {
+	if inspection == nil {
+		return nil
+	}
+	inputs := make([]PromptInputDefinition, len(inspection.Inputs))
+	for i, input := range inspection.Inputs {
+		inputs[i] = PromptInputDefinition{
+			Name:        input.Name,
+			Required:    input.Required,
+			ContentType: input.ContentType,
+			Description: input.Description,
+		}
+	}
+	return &PromptInspection{
+		PromptID:         inspection.PromptID,
+		PromptVersion:    inspection.PromptVersion,
+		PromptHash:       inspection.PromptHash,
+		DefaultProfileID: inspection.DefaultProfileID,
+		Inputs:           inputs,
+		OutputContract:   fromDomainOutputContract(inspection.OutputContract),
+	}
+}
+
 func fromDomainExecutionTargetPresence(presence domain.ExecutionTargetPresence) ExecutionTargetPresence {
 	return ExecutionTargetPresence{
 		Temperature:    presence.Temperature,

doc.go

@@ -3,23 +3,28 @@
 //
 // Applications construct an [Engine] with [NewEngine], select filesystem or
 // in-memory sources and optional engine-scoped [Backend] registrations, and
-// call [Engine.Prepare] or [Engine.Run]. Concrete registries, repositories,
-// validators, and the built-in OpenAI-compatible client remain internal
-// implementation details.
+// call [Engine.InspectPrompt], [Engine.InspectProfile], [Engine.Prepare],
+// [Engine.PrepareExecution], [Engine.Run], or [Engine.RunPrepared]. Concrete
+// registries, repositories, validators, and the built-in OpenAI-compatible
+// client remain internal implementation details.
 //
 // # Concurrency and ownership
 //
-// An Engine supports concurrent Prepare and Run calls. Engine-local backend
-// policies bound admitted Run calls and model generations where configured,
-// while different backend pools and unlimited backends continue independently.
-// An injected [LLMClient] or [ArtifactReader] can therefore still receive
-// concurrent calls and must be safe for that use.
+// An Engine supports concurrent InspectPrompt, InspectProfile, Prepare,
+// PrepareExecution, Run, and RunPrepared calls. Engine-local backend policies
+// bound admitted Run and RunPrepared calls and model generations where
+// configured, while different backend pools and unlimited backends continue
+// independently. An injected [LLMClient] or [ArtifactReader] can therefore
+// still receive concurrent calls and must be safe for that use.
 //
-// NewEngine copies in-memory profiles and backend definitions. Prepare and Run
-// copy request maps, slices, pointer values, and JSON-compatible extra
-// parameters before using them. Returned values and values passed to extension
-// interfaces are likewise isolated from engine state. Callers own those copies
-// and may mutate them after the call that supplied or returned them.
+// NewEngine copies in-memory profiles and backend definitions. Prepare,
+// PrepareExecution, and Run copy request maps, slices, pointer values, and
+// JSON-compatible extra parameters before using them. InspectPrompt and
+// InspectProfile return copied inspection values. Returned values and values
+// passed to extension interfaces are likewise isolated from engine state.
+// Callers own those copies and may mutate them after the call that supplied or
+// returned them. Returned structured errors are likewise caller-owned and may
+// be mutated without affecting engine state or another error.
 //
 // # Security and sensitive data
 //
@@ -45,10 +50,12 @@
 // [GenerateResponse], [ExecutionTargetPresence], and the string value types
 // used by those values.
 //
-// Construction values, including [Config], [Backend], [RunRequest],
-// [ArtifactRef], [ExecutionTargetOverride], [Profile], and
-// [OpenAICompatibleProfileConfig], do not have stable JSON representations.
-// Direct API keys are nevertheless excluded from JSON for every public value.
+// Construction, inspection, handle, and error values, including [Config],
+// [Backend], [RunRequest], [ArtifactRef], [ExecutionTargetOverride], [Profile],
+// [OpenAICompatibleProfileConfig], [ProfileInspection],
+// [PromptInputDefinition], [PromptInspection], [PreparedExecution], and
+// [CapacityError], do not have stable JSON representations. Direct API keys
+// are nevertheless excluded from JSON for every public value.
 //
 // JSON timestamps use time.Time's RFC 3339 encoding and are omitted when zero.
 // PreparedRun and RunResult durations are encoded as integer milliseconds in

docs/consumers/pkg-promptkit.md

@@ -40,11 +40,67 @@ validation, and default transport behavior. Source discovery, format
 validation, and profile precedence are defined by the
 [framework format reference](../formats.md).
 
+## Supply Embedded Application Defaults
+
+Use `WithFallbackProfileFS` when an application packages profile definitions
+that should apply unless an operator provides an ordinary configured profile
+with the same ID. For example, an application can embed its defaults while
+continuing to use `ProfileDir` for operator overrides:
+
+```go
+//go:embed profiles/*.yaml
+var applicationProfiles embed.FS
+
+engine, err := promptkit.NewEngine(promptkit.Config{
+	PromptDir:  "prompts",
+	ProfileDir: operatorProfileDir,
+},
+	promptkit.WithFallbackProfileFS(applicationProfiles, "profiles"),
+)
+```
+
+Keep application-owned profile IDs and definitions in the embedded source.
+Use the ordinary configured profile source for operator overrides. Leave
+`operatorProfileDir` empty when the operator did not configure an override
+directory; a non-empty path names an authoritative higher-precedence source,
+so an unavailable or unreadable directory is an error rather than a reason to
+fall back. The
+[framework format reference](../formats.md#source-and-profile-precedence)
+owns the exact profile format and lookup order; the
+[`WithFallbackProfileFS` GoDoc](../../engine.go) owns its option contract and
+validation rules.
+
+## Inspect A Prompt Before Preparation
+
+Use [`Engine.InspectPrompt`](../../engine.go) to check one configured prompt's
+declared inputs and output workflow without creating placeholder inputs or
+resolving a profile:
+
+```go
+inspection, err := engine.InspectPrompt(ctx, "meeting.summary", "")
+if err != nil {
+	return err
+}
+
+for _, input := range inspection.Inputs {
+	// Compare the declared input with application configuration.
+}
+```
+
+Use this configuration-time boundary when the application needs only the
+declared prompt interface. Use `InspectProfile` separately when it must also
+check a configured profile. Use `Prepare` when it needs inputs, schemas, or
+rendered messages, and use prepared execution when that work must remain tied
+to later execution. The method's [GoDoc](../../engine.go) owns exact fields,
+hash, ownership, and error semantics.
+
 ## Prepare Without Model Execution
 
 [`Engine.Prepare`](../../engine.go) resolves the selected prompt and profile,
 loads inputs and any structured-output schema, and renders messages without
-calling a model client:
+calling a model client. Choose it when the prepared value is the final
+inspection or persistence result and no later execution must be tied to that
+exact snapshot:
 
 ```go
 prepared, err := engine.Prepare(ctx, promptkit.RunRequest{
@@ -61,12 +117,48 @@ shows a complete runnable setup with a prompt file, in-memory profile, and
 inline input. Exact request requirements and prepared-result fields belong to
 the [`RunRequest` and `PreparedRun` GoDoc](../../types.go).
 
+## Prepare Now And Execute The Same Snapshot Later
+
+Use [`Engine.PrepareExecution`](../../engine.go) when an application must
+inspect or persist preflight details before deciding whether to start model
+work, while ensuring that later execution uses those exact rendered messages,
+inputs, target settings, and validation resources:
+
+```go
+preparedExecution, err := engine.PrepareExecution(ctx, promptkit.RunRequest{
+	PromptID: "meeting.summary",
+	Inputs: map[string]promptkit.ArtifactRef{
+		"note": promptkit.Inline("Synthetic meeting notes"),
+	},
+})
+if err != nil {
+	return err
+}
+defer preparedExecution.Discard()
+
+details := preparedExecution.Details()
+// Inspect or persist an application-selected safe subset of details.
+
+result, err := engine.RunPrepared(ctx, preparedExecution)
+```
+
+Preparation does not call the model or reserve backend capacity.
+`RunPrepared` executes from the retained snapshot rather than reloading
+consumer sources. The handle is opaque in-process state, while `Details`
+contains rendered content and remains subject to the application's data
+handling policy. The
+[`PreparedExecution` and method GoDoc](../../prepared_execution.go) and
+[engine operation GoDoc](../../engine.go) own exact lifecycle, engine-binding,
+credential, cancellation, timing, and error semantics.
+
 ## Execute And Validate
 
 [`Engine.Run`](../../engine.go) performs the same preparation, invokes the
 configured model client, classifies the generated artifact, and validates the
-content. A completed content check may return `ValidationFailed` in the result;
-an operational inability to validate returns an error.
+content in one call. Choose it when the application does not need a preflight
+boundary tied to the eventual execution. A completed content check may return
+`ValidationFailed` in the result; an operational inability to validate returns
+an error.
 
 The maintained
 [offline execution example](../../examples/go-library/run/main.go) injects a
@@ -90,13 +182,41 @@ replace execution settings or the complete output contract.
 The [public value GoDoc](../../types.go) defines nil, empty, zero, replacement,
 copy, and credential behavior. The
 [framework format reference](../formats.md) defines how those request values
-interact with prompt definitions, file-backed profiles, built-ins, schemas,
-and framework defaults.
+interact with prompt definitions, file-backed and application fallback
+profiles, built-ins, schemas, and framework defaults.
 
 For programmatic profiles,
 [`OpenAICompatibleProfile`](../../profiles.go) converts ordinary
 OpenAI-compatible settings into a value accepted by `WithProfiles`.
 
+### Inspect A Profile Before Prompt Work
+
+Use [`Engine.InspectProfile`](../../engine.go) to validate one configured
+profile without constructing a synthetic prompt or placeholder inputs. It
+resolves the profile's effective target but does not prepare or execute a
+prompt:
+
+```go
+inspection, err := engine.InspectProfile(ctx, profileID)
+if err != nil {
+	return err
+}
+
+target := inspection.EffectiveModelParams
+if target.APIKeyEnv != "" {
+	// Apply application policy for the named environment variable.
+} else if inspection.APIKeyRequired {
+	// Arrange a direct credential before later execution.
+}
+```
+
+Use this configuration-time boundary when only the profile and its target need
+checking. Use `Prepare` when the application also needs prompt, input, schema,
+or rendering work; use prepared execution when that work must remain tied to a
+later execution. Inspection reports credential requirements but leaves the
+timing of credential enforcement to the application. The method's
+[GoDoc](../../engine.go) owns its exact result and error contract.
+
 ### Set A Per-Run Session And Reasoning
 
 Supply a direct session ID when one prompt should be correlated with a
@@ -124,35 +244,86 @@ providers. The
 [`RunRequest` and `ExecutionTargetOverride` GoDoc](../../types.go) owns the
 exact normalization, precedence, error, copying, and exposure contract.
 
-### Register A Custom Backend
+### Configure A Local OpenAI-Compatible Endpoint
 
-Register a reusable OpenAI-compatible connection once, then select it from a
-profile. This local backend limits model generation to two simultaneous calls;
-because `QueueCapacity` is omitted, the engine admits up to 1024 additional
-calls waiting behind them:
+Choose the smallest configuration that fits how the endpoint will be reused.
+
+#### Use An Endpoint-Only Profile
+
+Put the endpoint directly on an in-memory profile when only that profile needs
+it and shared backend identity or capacity policy is unnecessary:
 
 ```go
 engine, err := promptkit.NewEngine(promptkit.Config{
 	PromptDir: "prompts",
 },
-	promptkit.WithBackend(promptkit.Backend{
-		ID:        "local",
-		Endpoint:  "http://localhost:8000/v1",
-		APIKeyEnv: "LOCAL_LLM_API_KEY",
-		ConcurrencyLimit: 2,
+	promptkit.WithProfiles(promptkit.Profile{
+		ID:       "local-summary",
+		Endpoint: "http://localhost:8000/v1",
+		Model:    "example-model",
 	}),
+)
+```
+
+Endpoint-only profiles have an empty backend ID and remain unrestricted by
+backend capacity policy.
+
+#### Use The Conventional Local Backend
+
+Use `LocalBackend` when profiles should share the conventional `local`
+identity, endpoint, and concurrency limit:
+
+```go
+engine, err := promptkit.NewEngine(promptkit.Config{
+	PromptDir: "prompts",
+},
+	promptkit.WithBackend(
+		promptkit.LocalBackend("http://localhost:8000/v1", 2),
+	),
 	promptkit.WithProfiles(promptkit.Profile{
 		ID:        "local-summary",
-		BackendID: "local",
+		BackendID: promptkit.BackendLocal,
 		Model:     "example-model",
 	}),
 )
 ```
 
+The helper is explicit: it does not pre-register a backend or read environment
+variables. Supplying a positive limit leaves queue capacity omitted, so normal
+backend registration selects the existing default waiting capacity of 1024.
+The returned value still enters the engine through `WithBackend`.
+
+#### Configure A Complete Backend
+
+Use a keyed `Backend` value for authentication, extra request parameters, an
+explicit queue capacity, a custom ID, or multiple local endpoints:
+
+```go
+noWaiting := 0
+engine, err := promptkit.NewEngine(promptkit.Config{
+	PromptDir: "prompts",
+},
+	promptkit.WithBackend(promptkit.Backend{
+		ID:               "local-gpu",
+		Endpoint:         "http://gpu-host:8000/v1",
+		APIKeyEnv:        "LOCAL_GPU_API_KEY",
+		ExtraParams:      map[string]any{"provider_option": "enabled"},
+		ConcurrencyLimit: 2,
+		QueueCapacity:    &noWaiting,
+	}),
+	promptkit.WithProfiles(promptkit.Profile{
+		ID:        "gpu-summary",
+		BackendID: "local-gpu",
+		Model:     "example-model",
+	}),
+)
+```
+
+Use distinct custom IDs when registering multiple local endpoints.
 Registrations belong to one engine and custom IDs cannot replace built-ins.
-The [`Backend` and `WithBackend` GoDoc](../../backends.go) defines validation,
-copying, uniqueness, exact concurrency-field semantics, and request-default
-behavior.
+The [`Backend`, `LocalBackend`, and `WithBackend` GoDoc](../../backends.go)
+defines exact construction, validation, copying, uniqueness, concurrency, and
+request-default behavior.
 
 Both file-backed and in-memory profiles select a registration through
 `backend` or `Profile.BackendID`. Profile and request endpoint overrides retain
@@ -173,8 +344,8 @@ zero:
 ```go
 noWaiting := 0
 backend := promptkit.Backend{
-	ID:               "local",
-	Endpoint:         "http://localhost:8000/v1",
+	ID:               "local-gpu",
+	Endpoint:         "http://gpu-host:8000/v1",
 	ConcurrencyLimit: 2,
 	QueueCapacity:    &noWaiting,
 }
@@ -236,6 +407,11 @@ When a limited backend has admitted all active and waiting calls, handle
 ```go
 result, err := engine.Run(ctx, request)
 if errors.Is(err, promptkit.ErrCapacityExceeded) {
+	var capacityErr *promptkit.CapacityError
+	if errors.As(err, &capacityErr) {
+		// Record capacityErr.BackendID using application-owned diagnostics.
+	}
+
 	// Apply application policy: shed work, report overload, or retry later.
 }
 ```
@@ -243,8 +419,9 @@ if errors.Is(err, promptkit.ErrCapacityExceeded) {
 A rejected call returns no partial result and does not invoke the model
 client. Promptkit does not prescribe retries or map this error to an HTTP
 status; those choices remain with the consuming application. The
-[`Engine.Run` and error GoDoc](../../engine.go) owns exact error and
-cancellation identities.
+[`CapacityError` GoDoc](../../capacity_error.go) owns the exact typed-error
+contract, while the [`Engine.Run` and error GoDoc](../../engine.go) owns broad
+error and cancellation identities.
 
 ## Application Boundary
 

docs/development.md

@@ -44,98 +44,3 @@ Start with:
 For cross-cutting changes, follow every applicable row. Do not create
 placeholder documents for packages, APIs, or integrations that do not yet
 exist.
-
-## Maintainer-Run Validation
-
-Promptkit does not currently use hosted CI. Maintainers are responsible for
-running the documented checks before accepting changes. Run the default Go
-validation from the Promptkit repository root:
-
-```sh
-go test ./...
-go test -race ./...
-go vet ./...
-go build ./...
-go run ./examples/go-library/prepare
-```
-
-Check formatting across every tracked Go file:
-
-```sh
-gofmt -l $(git ls-files '*.go')
-```
-
-The formatting command must produce no paths. Follow every added or changed
-Markdown link and confirm its target exists. Finally, check whitespace:
-
-```sh
-git diff --check
-```
-
-Documentation-only work does not require unrelated new tests, but it still
-requires link validation and `git diff --check`. Run the Go validation whenever
-documentation changes commands, examples, generated output, or another
-behavior checked by the module.
-
-## Focused Validation
-
-Use focused checks while iterating, then run the complete validation sequence
-before accepting the change. The root package supports:
-
-```sh
-go test .
-go vet .
-go build .
-```
-
-Filter tests by name without assuming a fixed internal package layout:
-
-```sh
-go test ./... -run 'TestName'
-```
-
-Replace `TestName` with a useful regular expression. Target only paths that
-exist, and consult the internal component overview for their owning
-documentation. A filtered or package-specific run does not replace the
-complete repository validation.
-
-## Coordinated Work With Scriptorium
-
-Promptkit and Scriptorium must remain independently valid. For temporary local
-integration, use either a Go workspace outside both repositories or an
-uncommitted replacement in the consuming module.
-
-If the repositories are sibling directories, run the workspace commands from
-their parent directory:
-
-```sh
-go work init ./promptkit ./scriptorium
-go work sync
-```
-
-Use the workspace only for coordinated local checks. From the same parent
-directory, remove it when finished:
-
-```sh
-rm -f go.work go.work.sum
-```
-
-Alternatively, from the Scriptorium repository root, temporarily point its
-Promptkit dependency at the sibling checkout:
-
-```sh
-go mod edit -replace gitea.maximumdirect.net/eric/promptkit=../promptkit
-```
-
-After coordinated checks, remove the replacement and reconcile module
-metadata:
-
-```sh
-go mod edit -dropreplace gitea.maximumdirect.net/eric/promptkit
-go mod tidy
-```
-
-Never commit `go.work`, `go.work.sum`, or a local filesystem `replace`
-directive. Before committing in either repository, inspect its module files and
-working tree independently. Published consumer versions must depend on a tagged
-Promptkit version, not a workspace, local replacement, or unpublished commit.

docs/formats.md

@@ -58,6 +58,11 @@ When a request omits a version, the selected prompt ID must identify exactly
 one definition. When it supplies a version, the ID and version pair must be
 unique.
 
+Exact prompt inspection uses this same configured source, strict decoding,
+referenced content-file resolution, and ID/version selection. It reports the
+selected definition's declared metadata without changing the prompt format or
+executing the definition.
+
 ### Inputs
 
 Each `inputs` item has these fields:
@@ -152,7 +157,7 @@ extra_params:
 | Field | Required | Meaning |
 | --- | --- | --- |
 | `id` | yes | Non-empty profile identifier. IDs must be unique within one source. |
-| `backend` | unless `endpoint` is present | Backend registry ID. It is trimmed and registry membership is checked when the profile is prepared. |
+| `backend` | unless `endpoint` is present | Backend registry ID. It is trimmed and registry membership is checked when the profile is prepared or inspected. |
 | `endpoint` | unless `backend` is present | Non-empty OpenAI-compatible base URL, including an API version path when required. When both connection fields are present, this overrides the backend endpoint without changing backend identity. |
 | `model` | yes | Non-empty provider model name. |
 | `temperature` | no | Number from 0 through 2. |
@@ -183,25 +188,27 @@ they also cannot collide with the standard fields listed in the
 
 Execution settings resolve in this order:
 
-1. framework defaults;
+1. the framework timeout baseline;
 2. the selected backend, when the profile names one;
 3. the selected profile; and
 4. request `ExecutionTargetOverride` values.
 
-The framework defaults are:
+The framework baseline is:
 
 | Setting | Default |
 | --- | --- |
-| `temperature` | `0` |
-| `max_tokens` | `0` |
-| `top_p` | `1` |
+| `temperature` | Unspecified and omitted from compatible provider requests unless a profile or runtime override selects it. |
+| `max_tokens` | Unspecified and omitted from compatible provider requests unless a profile or runtime override selects it. |
+| `top_p` | Unspecified and omitted from compatible provider requests unless a profile or runtime override selects it. |
 | `timeout_seconds` | `600` |
 
-Numeric zero in a file or in-memory profile means that the profile does not
-replace the framework default. Numeric request overrides use pointers, so an
-explicit zero is preserved. In particular, an explicit request
-`timeout_seconds` of zero disables the per-generation deadline while leaving
-the caller context and transport timeout intact.
+Numeric zero in a file or in-memory profile does not select a numeric value.
+For `temperature`, `max_tokens`, and `top_p`, it leaves the provider control
+unspecified. For `timeout_seconds`, it retains the framework deadline. Numeric
+request overrides use pointers, so an explicit zero is retained and sent to
+compatible providers. In particular, an explicit request `timeout_seconds` of
+zero disables the per-generation deadline while leaving the caller context and
+transport timeout intact.
 
 Non-empty profile strings replace backend defaults, and non-empty request
 strings replace both. Request reasoning is the exception: a nil
@@ -219,18 +226,25 @@ defines how the effective settings are serialized.
 ### Source And Profile Precedence
 
 An explicit request profile ID takes precedence over the prompt's
-`default_profile`. If neither is present, preparation fails.
+`default_profile`. If neither is present, preparation fails. Exact profile
+inspection instead takes one explicit profile ID and does not use a prompt
+default.
 
 Profile sources resolve matching IDs in this order:
 
 1. in-memory profiles supplied with `WithProfiles`;
-2. a profile file, `fs.FS`, or configured profile directory; and
-3. embedded built-in profiles.
+2. the ordinary configured source selected by a profile file, `fs.FS`, or
+   configured profile directory;
+3. application fallback profiles supplied with `WithFallbackProfileFS`; and
+4. embedded built-in profiles.
 
-A higher-precedence source falls back only when the profile is absent. An
-invalid matching profile is an error and does not fall back. In-memory
-`Profile` values follow the same ranges as YAML profiles. They use
-`APIKeyRequired` for request-scoped credentials instead of `api_key_env`.
+A profile source supplies a complete definition; definitions and their fields
+are not merged across sources. A higher-precedence source falls back only when
+the requested profile ID is absent. An invalid matching profile is an error and
+does not fall back. In-memory `Profile` values follow the same ranges as YAML
+profiles. They use `APIKeyRequired` for request-scoped credentials instead of
+`api_key_env`. Preparation and exact profile inspection use this same source
+precedence.
 
 ## Built-In Profile Catalog
 
@@ -238,8 +252,8 @@ Every built-in selects the `openrouter` backend. The engine's built-in backend
 registry supplies `https://openrouter.ai/api/v1` and the environment-variable
 name `OPENROUTER_API_KEY`, so individual profiles contain only model and
 generation settings. Built-in profile files do not repeat those connection
-values. A custom or in-memory profile with the same profile ID takes
-precedence.
+values. A configured, application fallback, or in-memory profile with the same
+profile ID takes precedence.
 
 | Provider | ID | Model |
 | --- | --- | --- |

docs/integrations/openai-compatible-chat.md

@@ -53,8 +53,9 @@ never also sent as a session header.
 
 The client conditionally includes:
 
-- `temperature`, `max_tokens`, and `top_p` when non-zero or explicitly
-  present;
+- `temperature`, `max_tokens`, and `top_p` only when selected by a profile or
+  runtime override, including an explicit runtime zero; they are absent when
+  unspecified;
 - non-empty `service_tier` and effective `reasoning_effort`; an explicitly
   disabled reasoning setting is empty and therefore omitted; and
 - `response_format` for JSON Schema structured output, including its name,

docs/internal/capacity.md

@@ -29,21 +29,28 @@ One pool owns immutable active and total limits plus mutex-protected admission
 count, active count, and ordered waiter list. Pool state exists only for the
 lifetime of its engine.
 
-## Bounded Run Admission
+## Bounded Execution Admission
 
-The runner asks the manager to admit a run after resolving the prompt, profile,
-selected backend, effective execution target, credentials, and output contract,
-but before schema loading, artifact loading, or rendering. Admission is
-immediate: a limited pool either reserves a slot or returns the internal
-`ErrCapacityExceeded` identity. The root facade maps that identity to the
-public error without treating it as an invalid request or generation failure.
+For ordinary `Run`, the runner asks the manager to admit after resolving the
+prompt, profile, selected backend, effective execution target, credentials, and
+output contract, but before schema loading, artifact loading, or rendering.
+`PrepareExecution` performs no admission. `RunPrepared` claims its handle,
+rechecks credential availability, and then asks the manager to admit the
+frozen backend before generation.
+
+Admission is immediate: a limited pool either reserves a slot or returns only
+the internal `ErrCapacityExceeded` identity. The runner attaches the selected
+backend identity at its use-case boundary, and the root facade translates that
+typed value without treating it as an invalid request or generation failure.
 
 The total admitted bound is the active-generation limit plus its configured
 waiting capacity. The returned release function is idempotent. The runner
-defers it as soon as admission succeeds and holds the lease across remaining
-preparation, initial generation, validation, every repair attempt, and all
-failure or cancellation exits. A repair is part of its original admission and
-does not reserve another bounded slot.
+defers it as soon as admission succeeds. An ordinary run holds the lease across
+remaining preparation, initial generation, validation, every repair attempt,
+and all failure or cancellation exits. Prepared execution holds the normal
+lease across generation, validation, every internal repair attempt, and all
+execution exits. A repair is part of its original admission and does not
+reserve another bounded slot.
 
 ## FIFO Generation Permits
 
@@ -90,11 +97,17 @@ unlimited admission. The
 FIFO transfer, canceled-waiter removal, grant/cancel races, independent pools,
 unlimited calls, passthrough behavior, and panic release.
 
-The [runner tests](../../internal/usecase/runner_test.go) own early admission,
-lease lifetime, failure release, and shared initial/repair scheduling. The
+The [runner tests](../../internal/usecase/runner_test.go) own ordinary early
+admission, lease lifetime, failure release, and shared initial/repair
+scheduling. The
+[prepared-execution use-case tests](../../internal/usecase/prepared_execution_test.go)
+own deferred admission, credential ordering, and prepared-execution lease
+release. The
 [external package capacity tests](../../capacity_contract_test.go) own the
 assembled public-engine behavior for configured limits, capacity errors,
 endpoint identity, engine independence, and injected clients. The
+[prepared-execution contract tests](../../prepared_execution_contract_test.go)
+own the public prepared-capacity boundary. The
 [root error-boundary tests](../../errors_internal_test.go) own preservation of
 the public generation category and context identity when generation is
 canceled.

docs/internal/llm.md

@@ -43,6 +43,21 @@ without making the model client depend on registry configuration.
 The implementation has no retry loop, tool-call support, provider catalog,
 inbound HTTP behavior, or durable session store.
 
+## Prepared Generation
+
+For [`RunPrepared`](../../engine.go), the runner supplies the model client with
+the target, rendered messages, and structured-output constraint retained by
+executable preparation. Execution does not reopen or rerender consumer
+sources.
+
+Before backend admission, the runner rechecks that the frozen credential
+environment-variable name is available. The handle does not retain the
+environment value; the model client resolves the value visible when generation
+begins. A direct request key remains in private execution state only until the
+claimed execution finishes or an unclaimed handle is discarded. Exact public
+ownership and redaction semantics belong to the
+[`PreparedExecution` GoDoc](../../prepared_execution.go).
+
 ## Failure Categories
 
 The package preserves distinct error identities for invalid client

docs/internal/overview.md

@@ -11,23 +11,23 @@ contributor workflow and validation.
 
 | Component | Implemented responsibility | References |
 | --- | --- | --- |
-| Root `promptkit` package | Provides the supported engine facade, source, backend-registration, and injection options, public request and result values, profile construction, extension interfaces, value conversion, redacted formatting, public error mapping, and engine-local assembly. | [Package GoDoc](../../doc.go), [backend API](../../backends.go), [engine assembly](../../engine.go) |
+| Root `promptkit` package | Provides the supported engine facade, source, backend-registration, and injection options, public request, result, prompt-inspection, and profile-inspection values, opaque prepared-execution handles, profile construction, extension interfaces, value conversion, redacted formatting, typed capacity errors, public error mapping, and engine-local profile-source assembly including application fallbacks. | [Package GoDoc](../../doc.go), [prepared execution](../../prepared_execution.go), [backend API](../../backends.go), [engine assembly](../../engine.go) |
 | `examples/go-library/prepare` | Demonstrates an offline downstream consumer using a prompt file, in-memory profile, inline input, and `Prepare`. It is not a public library package. | [Example program](../../examples/go-library/prepare/main.go) |
 | `examples/go-library/run` | Demonstrates an offline downstream consumer using a prompt file, in-memory profile, inline input, an injected deterministic model client, and `Run`. It is not a public library package. | [Example program](../../examples/go-library/run/main.go) |
 | `internal/backend` | Constructs each engine's immutable registry from the built-in OpenRouter definition and consumer additions, validates and defensively copies definitions through the shared JSON-value package, and consumes the LLM-owned OpenAI-compatible reserved request-field rule. | [Backend registry](../../internal/backend/registry.go) |
-| `internal/capacity` | Owns engine-local bounded run admission and FIFO model-generation permits for limited backend IDs, including cancellation-safe waiter removal and client wrapping. | [Internal capacity management](capacity.md) |
+| `internal/capacity` | Owns engine-local bounded execution admission and FIFO model-generation permits for limited backend IDs, including cancellation-safe waiter removal and client wrapping. | [Internal capacity management](capacity.md) |
 | `internal/domain` | Defines internal framework values for requests, artifacts, prompt definitions, profiles, execution targets, rendering, generation, and validation. | [Domain declarations](../../internal/domain/domain.go) |
 | `internal/defaults` | Defines application-neutral framework constants and constructs the default execution target. It contains no CLI, server, or inbound HTTP limits. | [Framework defaults](../../internal/defaults/defaults.go) |
 | `internal/filecatalog` | Provides deterministic YAML discovery and path helpers for operating-system filesystems and `fs.FS` sources. | [File catalog](../../internal/filecatalog/catalog.go) |
-| `internal/jsonvalue` | Validates and deeply copies JSON-compatible extra-parameter trees while preserving supported concrete value types. | [JSON values](../../internal/jsonvalue/jsonvalue.go) |
+| `internal/jsonvalue` | Validates and deeply copies JSON-compatible extra-parameter and prepared-schema trees while preserving supported concrete value types. | [JSON values](../../internal/jsonvalue/jsonvalue.go) |
 | `internal/promptdef` | Loads strictly decoded, validated prompt definitions from filesystem and `fs.FS` sources, including version selection and contained file-backed message content. | [Framework formats](../formats.md), [prompt-definition repository](../../internal/promptdef/filesystem_repository.go) |
 | `internal/profile` | Loads strictly decoded, validated execution profiles, including backend selection, from filesystem and `fs.FS` sources and composes repositories with error-preserving fallback. | [Framework formats](../formats.md), [profile repositories](../../internal/profile/filesystem_repository.go) |
-| `internal/profile/builtin` | Embeds the built-in profile catalog, whose entries select OpenRouter, and combines it with an optional primary repository. | [Built-in catalog](../formats.md#built-in-profile-catalog), [repository](../../internal/profile/builtin/repository.go) |
+| `internal/profile/builtin` | Embeds the built-in profile catalog, whose entries select OpenRouter. | [Built-in catalog](../formats.md#built-in-profile-catalog), [repository](../../internal/profile/builtin/repository.go) |
 | `internal/prompt` | Renders prompt messages from Go templates with artifact, variable, session, and cache-control data. | [Go-template renderer](../../internal/prompt/go_renderer.go) |
 | `internal/artifact` | Resolves ordinary inline and unrestricted caller-selected file references into copied artifacts with metadata and hashes. | [Internal sources and validation](sources.md) |
-| `internal/validate` | Validates basic, JSON, and JSON Schema output using operating-system filesystem or `fs.FS` schema sources. | [Framework formats](../formats.md#schemas), [internal sources and validation](sources.md) |
+| `internal/validate` | Validates basic, JSON, and JSON Schema output using operating-system filesystem or `fs.FS` schema sources and creates frozen validation plans for prepared execution. | [Framework formats](../formats.md#schemas), [internal sources and validation](sources.md) |
 | `internal/llm` | Defines the internal generation boundary and implements outbound OpenAI-compatible chat requests from resolved execution targets, including response decoding, authentication, deadline handling, and ownership of the OpenAI-compatible reserved request-field policy. | [Internal model client](llm.md) |
-| `internal/usecase` | Resolves backend, profile, and request settings and coordinates preparation and execution across internal sources, rendering, artifact loading, generation, validation, and optional repair. | [Internal runner](runner.md) |
+| `internal/usecase` | Resolves prompt definitions and hashes, profiles, backends, and targets for exact inspection and request settings for preparation, and coordinates ordinary execution and one-attempt prepared execution across internal sources, rendering, artifact loading, generation, validation, capacity, and optional repair. | [Internal runner](runner.md), [prepared-execution implementation](../../internal/usecase/prepared_execution.go) |
 
 The root package assembles these internal components without exposing their
 representations. Consumers depend only on the root facade.

docs/internal/runner.md

@@ -28,6 +28,34 @@ constructor does not enable one.
 Each invocation carries its state in request, prepared-run, and result values.
 The runner has no durable run or session store.
 
+## Shared Prompt Selection
+
+The runner uses one prompt-selection and hashing boundary for ordinary
+preparation and exact prompt inspection. Preparation retains its early
+request-ID check before direct-session normalization; both operations then use
+the configured prompt repository to select one definition, load referenced
+message content, and calculate the same prompt hash.
+
+Inspection stops after that structural lookup. It does not parse templates or
+touch profile, artifact, schema, renderer, validator, admission, or model
+collaborators. The root [`Engine.InspectPrompt`](../../engine.go) GoDoc owns
+the public operation's exact contract.
+
+## Shared Profile Selection
+
+The runner uses one profile-selection and target-resolution boundary for
+ordinary preparation and exact profile inspection. Preparation first selects a
+request profile or a prompt default; inspection begins with its required
+explicit profile ID. Both then apply the ordinary source precedence, resolve a
+named backend, and construct the effective target from framework, backend, and
+profile values.
+
+Inspection stops after the resulting endpoint and model are structurally
+validated. It does not check credential availability or perform prompt,
+artifact, schema, rendering, admission, or model-client work. The root
+[`Engine.InspectProfile`](../../engine.go) GoDoc owns the public operation's
+exact contract.
+
 ## Shared Preparation Pipeline
 
 `Prepare` and `Run` share one private preparation pipeline split at the point
@@ -124,15 +152,17 @@ validation failures. Wrapping preserves the package identities mapped by the
 public facade and retains collaborator identities where they are part of the
 internal contract.
 
-Admission capacity exhaustion retains the internal capacity identity and adds
-the selected backend ID as context. It is not recategorized as an invalid
-request or generation failure, and no partial result is returned. A context
-already done at admission retains its context identity directly. Cancellation
-while waiting for an active generation permit prevents client invocation when
-it wins the grant race; the model-client boundary then preserves the context
-error through the generation-failure category. Deferred release restores the
-admission lease on preparation, generation, validation, repair, and
-cancellation failures.
+Admission capacity exhaustion retains the internal capacity identity. At the
+use-case boundary, the runner attaches the selected backend ID in an internal
+typed error, and the root facade copies that value into the public
+[`CapacityError`](../../capacity_error.go) without parsing diagnostic text. It
+is not recategorized as an invalid request or generation failure, and no
+partial result is returned. A context already done at admission retains its
+context identity directly. Cancellation while waiting for an active generation
+permit prevents client invocation when it wins the grant race; the model-client
+boundary then preserves the context error through the generation-failure
+category. Deferred release restores the admission lease on preparation,
+generation, validation, repair, and cancellation failures.
 
 Other context cancellation propagates through the invoked collaborator and is
 classified by the owning operation.

docs/internal/sources.md

@@ -16,6 +16,11 @@ validation modes, built-in catalog, and source precedence.
 definitions, selects an ID and optional version, and resolves file-backed
 message content within the selected operating-system or `fs.FS` source.
 
+Exact prompt inspection performs one point-in-time lookup through that same
+repository and validates referenced message content before returning declared
+metadata. It does not parse templates or read profile, input, or schema
+sources, and it does not retain the definition for a later execution.
+
 Its package tests own prompt selection, strict decoding, definition validation,
 duplicate detection, and source containment:
 [prompt-definition repository tests](../../internal/promptdef/repository_test.go).
@@ -23,21 +28,30 @@ duplicate detection, and source containment:
 ## Profiles And Built-Ins
 
 `internal/profile` loads and validates execution profiles from an
-operating-system filesystem or an `fs.FS`. It supports a primary repository
-with fallback only when the primary reports that a profile is absent. Strict
-YAML decoding recognizes the optional `backend` field, trims its value, and
-requires a model plus at least one non-blank backend or endpoint. Loading does
-not check registry membership because the available registry belongs to the
-assembled engine; the runner checks membership during preparation.
+operating-system filesystem or an `fs.FS`. Its overlay repository consults the
+next repository only when the higher-precedence repository reports that a
+profile is absent. Strict YAML decoding recognizes the optional `backend`
+field, trims its value, and requires a model plus at least one non-blank
+backend or endpoint. Loading does not check registry membership because the
+available registry belongs to the assembled engine; the runner checks
+membership during preparation and exact profile inspection.
 
-`internal/profile/builtin` embeds the maintained built-in profile catalog and
-can place a caller-selected repository ahead of that catalog. Every embedded
-profile selects `openrouter` and inherits its endpoint and credential
-environment-variable name from the built-in backend registry rather than
-repeating those values. Profile behavior is owned by the
-[profile repository tests](../../internal/profile/repository_test.go), while
-catalog completeness, the backend-selection invariant, duplicate IDs, and
-overlay behavior are owned by the
+The root engine assembles profile repositories in precedence order: in-memory
+profiles, one ordinary configured source, an application fallback source, then
+the embedded built-in catalog. An explicit file or `fs.FS` profile source
+replaces `Config.ProfileDir` within the ordinary configured-source category.
+
+Exact profile inspection performs one point-in-time lookup through those
+profile sources and checks the resolved target without reading prompt, input,
+or schema sources. It does not retain that lookup for a later execution.
+
+`internal/profile/builtin` embeds the maintained built-in profile catalog.
+Every embedded profile selects `openrouter` and inherits its endpoint and
+credential environment-variable name from the built-in backend registry rather
+than repeating those values. Profile loading and overlay behavior are owned by
+the [profile repository tests](../../internal/profile/repository_test.go),
+while catalog completeness, the backend-selection invariant, and duplicate IDs
+are owned by the
 [built-in repository tests](../../internal/profile/builtin/repository_test.go).
 
 ## Ordinary Artifacts
@@ -68,6 +82,22 @@ filesystem or an `fs.FS`. Invalid generated content is returned as a validation
 result; inability to load, register, or compile a schema is an operational
 error.
 
+For executable preparation, the built-in validators create a frozen validation
+plan. None, basic, and JSON modes retain the effective output contract without
+source access. JSON Schema mode loads the root document, resolves and compiles
+every transitive reference during preparation, and retains the compiled
+validator. The provider-facing structured-output metadata uses that same
+captured root document.
+
+`PrepareExecution` also completes prompt and profile selection, artifact
+loading and hashing, session and message rendering, and target resolution.
+`RunPrepared` uses the retained source-derived state and validation plan; it
+does not reopen prompt, profile, input, or schema sources and does not rerender
+the request. By contrast, ordinary `Prepare` produces a preparation value only:
+a later `Run` performs its own source resolution and preparation.
+
 The [validator tests](../../internal/validate/standard_validator_test.go) own
-basic, JSON, JSON Schema, source resolution, schema loading, compilation, and
-content-failure behavior.
+basic, JSON, JSON Schema, source resolution, schema loading, compilation,
+frozen-reference behavior, and content-failure behavior. Prepared execution
+orchestration is owned by the
+[use-case tests](../../internal/usecase/prepared_execution_test.go).

docs/releases/v0.2.0.md

@@ -90,7 +90,7 @@ engine, err := promptkit.NewEngine(
 ```
 
 See the
-[custom-backend consumer guide](../consumers/pkg-promptkit.md#register-a-custom-backend)
+[local-endpoint consumer guide](../consumers/pkg-promptkit.md#configure-a-local-openai-compatible-endpoint)
 for task-oriented usage. The
 [`Backend` and `WithBackend` GoDoc](../../backends.go) owns exact registration,
 validation, copying, defaulting, and uniqueness semantics. The

docs/releases/v0.3.0.md

@@ -0,0 +1,74 @@
+# Promptkit v0.3.0
+
+This supplemental changelog summarizes the consumer-facing changes from
+`v0.2.0` to `v0.3.0`. The annotated `v0.3.0` tag is the authoritative release
+record. Exact current contracts belong to the linked GoDoc and durable
+documentation.
+
+## Summary
+
+`v0.3.0` adds a concise way to register the common local OpenAI-compatible
+backend configuration:
+
+- `BackendLocal` provides the conventional, non-reserved backend ID `"local"`;
+  and
+- `LocalBackend` constructs an ordinary `Backend` from an endpoint and
+  concurrency limit.
+
+The helper is explicit and additive. It does not pre-register a backend, read
+environment variables, select a model, or replace the complete `Backend`
+configuration interface.
+
+## Compatibility
+
+Existing `v0.2.0` consumers require no migration. Endpoint-only profiles,
+complete custom `Backend` values, the built-in OpenRouter backend, and existing
+registrations using the literal ID `"local"` continue to work unchanged.
+
+## Upgrade
+
+Update the module dependency with:
+
+```sh
+go get gitea.maximumdirect.net/eric/promptkit@v0.3.0
+go mod tidy
+```
+
+Run the consuming project's ordinary tests and race-enabled tests after the
+upgrade.
+
+## Configure A Local Backend
+
+Register the convenience value through the existing `WithBackend` option and
+select it from one or more profiles:
+
+```go
+engine, err := promptkit.NewEngine(
+	promptkit.Config{PromptDir: "prompts"},
+	promptkit.WithBackend(
+		promptkit.LocalBackend("http://localhost:8000/v1", 2),
+	),
+	promptkit.WithProfiles(promptkit.Profile{
+		ID:        "local-summary",
+		BackendID: promptkit.BackendLocal,
+		Model:     "example-model",
+	}),
+)
+```
+
+Use an endpoint-only profile when shared backend identity and capacity policy
+are unnecessary. Continue to use a complete keyed `Backend` value for custom
+IDs, authentication, extra request parameters, explicit queue capacity, or
+multiple local endpoints.
+
+See the
+[local-endpoint consumer guide](../consumers/pkg-promptkit.md#configure-a-local-openai-compatible-endpoint)
+for task-oriented configuration choices. The
+[`BackendLocal`, `LocalBackend`, and `WithBackend` GoDoc](../../backends.go)
+owns their exact construction, registration, validation, and concurrency
+semantics.
+
+## Consumer Action
+
+None. Adopt the convenience constructor when it simplifies local endpoint
+configuration.

docs/releases/v0.4.0.md

@@ -0,0 +1,189 @@
+# Promptkit v0.4.0
+
+This supplemental changelog and adoption guide summarizes the consumer-facing
+changes from `v0.3.0` to `v0.4.0`. The annotated `v0.4.0` tag is the
+authoritative release record. Exact current contracts belong to the linked
+GoDoc and durable documentation.
+
+## Summary
+
+`v0.4.0` adds four complementary capabilities:
+
+- opaque prepared-execution handles for preparing once, inspecting safe
+  details, and executing the same frozen snapshot;
+- exact prompt-definition inspection without profile resolution or execution;
+- exact profile inspection without selecting a prompt or checking credential
+  availability; and
+- structured backend identity on engine admission-capacity rejection.
+
+These APIs let consumers perform more precise preflight work and retain useful
+operational context without reproducing Promptkit's internal resolution logic.
+
+## Compatibility
+
+The release is additive for `v0.3.0` consumers. Existing uses of `Prepare`,
+`Run`, backend registration, endpoint-only profiles, local-backend helpers,
+runtime overrides, public JSON values, and error sentinels continue to work
+without migration.
+
+Capacity rejection now returns a structured error while continuing to match
+`ErrCapacityExceeded` through `errors.Is`. Error-string wording and direct
+sentinel equality were not public contracts.
+
+The new inspection values, capacity error, and prepared-execution handle do not
+have stable JSON representations. `PreparedExecution.Details` returns the
+existing stable `PreparedRun` value.
+
+## Upgrade
+
+Update the module dependency with:
+
+```sh
+go get gitea.maximumdirect.net/eric/promptkit@v0.4.0
+go mod tidy
+```
+
+Run the consuming project's ordinary and race-enabled tests after upgrading.
+No source migration is required.
+
+## Prepare Once And Execute The Same Snapshot
+
+Consumers that need to persist preparation details before generation can now
+prepare an opaque, engine-bound execution:
+
+```go
+prepared, err := engine.PrepareExecution(ctx, request)
+if err != nil {
+	// Handle preparation failure.
+}
+defer prepared.Discard()
+
+details := prepared.Details()
+// Persist a consumer-selected, appropriately protected preparation record.
+
+result, err := engine.RunPrepared(ctx, prepared)
+```
+
+Preparation freezes the selected sources, rendered messages, effective
+settings, input content, structured-output metadata, and validation resources
+needed by execution. `Details` returns a fresh, caller-owned,
+credential-redacted `PreparedRun`.
+
+A handle belongs to its creating engine and permits one execution attempt.
+`RunPrepared` consumes that attempt on success and on operational failure.
+`Discard` is idempotent and releases an unclaimed handle's execution-only
+state. Consumers should discard handles they will not execute, particularly
+when a direct request API key may be retained privately until claim or
+discard.
+
+Prepared execution does not reserve backend admission during preparation.
+Credential availability and backend admission are checked when execution
+begins. The execution context is independent of the preparation context.
+
+See the
+[prepared-execution consumer guide](../consumers/pkg-promptkit.md#prepare-now-and-execute-the-same-snapshot-later),
+the [`PreparedExecution` GoDoc](../../prepared_execution.go), and the
+[`Engine.PrepareExecution` and `Engine.RunPrepared` GoDoc](../../engine.go)
+for the exact lifecycle, ownership, cancellation, capacity, timing, and
+failure contracts.
+
+## Inspect A Prompt
+
+`Engine.InspectPrompt` resolves one prompt ID and optional version through the
+engine's configured prompt source:
+
+```go
+inspection, err := engine.InspectPrompt(ctx, "report.summary", "")
+```
+
+The result includes prompt identity, the opaque prompt hash, declared default
+profile ID, declared input metadata, and normalized output contract. It
+structurally loads the selected definition and referenced message content but
+does not resolve a profile, load schemas or artifacts, render templates,
+reserve capacity, or contact a model.
+
+Use inspection for exact configuration checks and metadata discovery. Use
+`PrepareExecution` rather than relying on a prior inspection when later
+execution must freeze one exact source state, because filesystem-backed
+inspection is only a point-in-time lookup.
+
+See the
+[prompt-inspection consumer guide](../consumers/pkg-promptkit.md#inspect-a-prompt-before-preparation)
+and [`Engine.InspectPrompt` GoDoc](../../engine.go) for exact selection,
+ownership, and error behavior.
+
+## Inspect A Profile
+
+`Engine.InspectProfile` resolves one explicit profile independently of a
+prompt:
+
+```go
+inspection, err := engine.InspectProfile(ctx, "report-production")
+```
+
+The result includes the resolved effective execution target and whether a
+later request must provide a direct credential. Environment-variable names may
+be reported, but inspection does not read credential values or require the
+named variable to be populated.
+
+Inspection applies the engine's profile source precedence and resolves any
+selected backend. It does not load a prompt, render content, reserve capacity,
+or contact a model.
+
+See the
+[profile-inspection consumer guide](../consumers/pkg-promptkit.md#inspect-a-profile-before-prompt-work)
+and [`Engine.InspectProfile` GoDoc](../../engine.go) for the exact resolution,
+credential, ownership, and error contracts.
+
+## Identify Capacity-Rejected Backends
+
+Calls rejected at Promptkit's bounded engine admission boundary continue to
+match `ErrCapacityExceeded`. Consumers can additionally obtain the selected
+registered backend ID without parsing diagnostic text:
+
+```go
+result, err := engine.Run(ctx, request)
+if errors.Is(err, promptkit.ErrCapacityExceeded) {
+	var capacityErr *promptkit.CapacityError
+	if errors.As(err, &capacityErr) {
+		// Record capacityErr.BackendID using application-owned diagnostics.
+	}
+
+	// Apply application-owned overload or retry policy.
+}
+```
+
+The structured error applies to `Run` and `RunPrepared` admission rejection.
+It does not represent provider throttling, quota exhaustion, cancellation
+while waiting for generation capacity, or another model-client failure.
+Promptkit does not prescribe retry timing or transport status mapping.
+
+See the
+[error-handling consumer guide](../consumers/pkg-promptkit.md#handle-errors),
+the [`CapacityError` GoDoc](../../capacity_error.go), and the
+[`ErrCapacityExceeded` GoDoc](../../engine.go) for the canonical contracts.
+
+## Public API Additions
+
+The release adds:
+
+- `Engine.PrepareExecution`;
+- `Engine.RunPrepared`;
+- `PreparedExecution`, including `Details`, `Discard`, `String`, and
+  `GoString`;
+- `Engine.InspectPrompt`;
+- `PromptInspection`;
+- `PromptInputDefinition`;
+- `Engine.InspectProfile`;
+- `ProfileInspection`; and
+- `CapacityError`.
+
+No public API was removed.
+
+## Consumer Action
+
+None. Existing `v0.3.0` workflows may upgrade without adopting the new APIs.
+
+Consumers that adopt prepared execution should discard unused handles.
+Consumers that need backend-specific capacity diagnostics may add an
+`errors.As` check while retaining their existing `errors.Is` classification.

docs/releases/v0.5.0.md

@@ -0,0 +1,125 @@
+# Promptkit v0.5.0
+
+This supplemental changelog and migration guide summarizes the consumer-facing
+changes from `v0.4.0` to `v0.5.0`. The annotated `v0.5.0` tag is the
+authoritative release record. Exact current contracts belong to the linked
+GoDoc and durable documentation.
+
+## Summary
+
+`v0.5.0` makes provider requests less prescriptive and adds an application
+fallback layer for profile definitions:
+
+- unset optional provider controls are omitted from OpenAI-compatible request
+  bodies instead of being populated with framework values; and
+- `WithFallbackProfileFS` lets an application package profile defaults that
+  operators can override through the existing ordinary profile sources.
+
+These changes let compatible providers apply their own model defaults while
+giving applications stable embedded profile IDs without weakening operator
+configuration precedence.
+
+## Compatibility
+
+The release adds one public function and removes no public declaration.
+Existing source code should continue to compile.
+
+There is one intentional behavior change: when no profile or runtime override
+selects `top_p`, Promptkit no longer sends the former framework value of `1`.
+It omits `top_p` and lets the provider choose its behavior. Unset
+`temperature` and `max_tokens` are likewise omitted. Explicit nonzero profile
+values and runtime values—including explicit runtime zero values—retain their
+precedence and wire effect.
+
+Consumers that relied on Promptkit always sending `top_p: 1` should add that
+value to the relevant profile or runtime override before upgrading. Consumers
+that did not rely on the implicit sampling value require no migration.
+
+Application fallback profiles are opt-in. Engines that do not call
+`WithFallbackProfileFS` retain the previous profile-source behavior.
+
+## Upgrade
+
+Update the module dependency with:
+
+```sh
+go get gitea.maximumdirect.net/eric/promptkit@v0.5.0
+go mod tidy
+```
+
+Run the consuming project's ordinary and race-enabled tests after upgrading.
+If request payloads or model behavior are asserted in fixtures, review them for
+the optional-parameter omission described below.
+
+## Omitted Optional Provider Controls
+
+The built-in OpenAI-compatible client now includes `temperature`,
+`max_tokens`, and `top_p` only when a profile or runtime override selects the
+value. An explicit runtime zero remains present because runtime override
+pointers distinguish zero from an unspecified value.
+
+Promptkit's positive generation deadline remains a framework concern and is
+not a provider request-body default. Required request fields, session IDs,
+structured output, reasoning selection, credentials, and explicit extra
+parameters retain their existing behavior.
+
+See the [framework default and precedence reference](../formats.md#defaults-and-overrides),
+the [`ExecutionTargetOverride` GoDoc](../../types.go), and the
+[OpenAI-compatible request-body contract](../integrations/openai-compatible-chat.md#request-body)
+for current details.
+
+## Embedded Application Fallback Profiles
+
+Applications can package ordinary profile YAML in an `fs.FS` and register it
+as a fallback source:
+
+```go
+//go:embed profiles/*.yaml
+var applicationProfiles embed.FS
+
+engine, err := promptkit.NewEngine(promptkit.Config{
+	PromptDir:  "prompts",
+	ProfileDir: operatorProfileDir,
+},
+	promptkit.WithFallbackProfileFS(applicationProfiles, "profiles"),
+)
+```
+
+Leave `operatorProfileDir` empty when no operator source is configured. A
+configured ordinary source is authoritative: a matching definition overrides
+the application fallback, while a read or validation failure remains an error
+instead of silently reaching a lower layer.
+
+Profile definitions resolve in this order:
+
+1. in-memory profiles supplied with `WithProfiles`;
+2. the ordinary configured source selected by `WithProfileFile`,
+   `WithProfileFS`, or `Config.ProfileDir`;
+3. the application source supplied with `WithFallbackProfileFS`; and
+4. Promptkit's embedded built-in profiles.
+
+Only an absent profile ID falls through. Sources provide complete profiles and
+do not merge fields. Loading remains lazy, and the new source uses the existing
+strict profile YAML and credential rules.
+
+See the
+[embedded-default consumer guidance](../consumers/pkg-promptkit.md#supply-embedded-application-defaults),
+the [`WithFallbackProfileFS` GoDoc](../../engine.go), and the
+[profile source reference](../formats.md#source-and-profile-precedence) for
+current details.
+
+## Public API Changes
+
+The release adds:
+
+- `WithFallbackProfileFS`.
+
+No public declaration was removed or changed.
+
+## Consumer Action
+
+- Review any workflow that depended on Promptkit's implicit `top_p: 1` and
+  configure the value explicitly when required.
+- Optionally adopt `WithFallbackProfileFS` when an application should package
+  overridable profile defaults.
+- Run consumer tests after updating the module dependency.

docs/roadmap/concurrency.md

@@ -1,236 +0,0 @@
-# Backend-Specific Concurrency Management
-
-**Status:** Complete.
-
-## Purpose
-
-This roadmap defines the scope and target end state for engine-local,
-backend-specific concurrency management. It records the intended capability,
-consumer value, and important policy choices.
-
-This document is planning material, not a description of current behavior.
-Current exported contracts remain owned by Go declarations and GoDoc, backend
-registration guidance by the
-[consumer guide](../consumers/pkg-promptkit.md#register-a-custom-backend), and
-implemented orchestration by the
-[internal runner document](../internal/runner.md).
-
-## Motivation
-
-Different model backends can sustain very different request loads. A local
-network endpoint may need a small concurrency limit, while OpenRouter can
-usually accept substantially more simultaneous work. Requiring every consumer
-to build its own semaphores and queues would duplicate routing knowledge,
-create inconsistent cancellation behavior, and make it easy for one caller to
-bypass the intended backend limit.
-
-Promptkit should own this coordination because it already resolves each run to
-an engine-scoped backend identity and owns every model-generation call made by
-the runner. Consumers should continue submitting ready-to-run requests through
-the synchronous API, including concurrently from multiple goroutines, without
-implementing their own backend scheduler.
-
-The buffered queue is a safety boundary, not an ordinary throughput
-restriction. Its primary purpose is to prevent a bug or unintended submission
-loop from creating an unbounded in-memory backlog.
-
-## Scope
-
-The feature will add optional concurrency policy to registered backends and
-coordinate `Run` calls against independent per-backend capacity pools.
-
-Each policy has two distinct controls:
-
-- an active-generation limit, which protects the backend from too many
-  simultaneous model requests; and
-- a bounded waiting capacity, which protects the process from admitting an
-  unbounded backlog.
-
-Concurrency policy belongs to a backend registration. It is not a profile
-model parameter and cannot be overridden per run. Profiles select the policy
-through their backend ID, while a profile or request endpoint override remains
-in the selected backend's pool.
-
-`Prepare` does not call a model and will remain outside concurrency admission.
-
-## Defaults And Configuration
-
-The built-in OpenRouter backend will use:
-
-- an active-generation limit of 16; and
-- a waiting capacity of 1024.
-
-The waiting default is intentionally generous. Reaching it should indicate
-abnormal submission pressure rather than normal application behavior.
-
-Consumer-registered backends will remain unlimited unless the consumer
-configures an active-generation limit. When a consumer enables a limit and
-does not specify waiting capacity, the waiting capacity will default to 1024.
-Consumers may configure a different bounded capacity, including zero when
-they want no admitted backlog beyond the active-limit-sized run set.
-
-The public representation must distinguish an omitted waiting capacity from
-an explicit zero.
-
-Endpoint-only profiles have no backend registration from which to obtain
-policy and will remain unlimited. A future engine-wide or endpoint-keyed
-policy can be considered separately if consumers demonstrate that need.
-
-Invalid limits or capacities will fail engine construction as invalid
-configuration. Policy values will be copied into engine-owned immutable state
-along with the rest of the backend registration.
-
-## Admission And Execution Behavior
-
-`Run` remains a synchronous, wait-for-result operation. Concurrent callers may
-block inside `Run` while waiting for their selected backend, then receive the
-ordinary result or error from that invocation.
-
-For a configured pool, the active-generation limit plus the waiting capacity
-defines the maximum number of concurrent `Run` invocations that Promptkit will
-accept for that backend. A waiting capacity of zero therefore accepts no more
-runs than the active limit. Admission is immediate: a call either reserves one
-of those bounded slots or receives the capacity error. An accepted run may
-then wait internally for active-generation capacity.
-
-For a limited backend, Promptkit will bound the number of accepted runs before
-expensive artifact loading, prompt rendering, and large defensive copies where
-practical. Lightweight prompt, profile, and backend resolution may occur first
-when it is required to identify the correct capacity pool. This pre-admission
-resolution must not become a second execution-precedence path with behavior
-that can drift from `Prepare`.
-
-An accepted run retains its admission until it completes or fails. Every
-actual model-generation call for that run must separately observe the
-backend's active-generation limit. This includes:
-
-- the initial generation;
-- every output-repair generation; and
-- calls made through either the built-in or an injected model client.
-
-Preparation and output validation should not hold an active-generation permit.
-A repair remains part of its already-admitted run, but reacquires active
-generation capacity so repairs cannot exceed the backend limit. It must not be
-rejected merely because new runs filled the waiting queue after its initial
-generation.
-
-Within one backend pool, waiting generation calls should be served in FIFO
-order, subject to canceled calls being removed. Different backend pools make
-progress independently; a saturated local backend must not consume
-OpenRouter's active or waiting capacity.
-
-The feature will not promise ordering across backend pools or completion order
-among admitted runs.
-
-## Capacity Failure And Cancellation
-
-When a backend's bounded waiting capacity is full, a new `Run` call will fail
-promptly rather than waiting outside the bounded admission system. The public
-API will expose a recognizable capacity-exhaustion error identity distinct
-from invalid configuration, invalid requests, and model-client failures.
-Rejected calls return no partial result and do not invoke the model client.
-
-Waiting within the admitted backlog or for active-generation capacity must
-honor the caller's context. Cancellation or deadline expiry will:
-
-- stop waiting promptly;
-- release any admission or generation capacity held by that invocation;
-- preserve the applicable context error identity; and
-- avoid invoking the model client if cancellation wins before generation
-  starts.
-
-Capacity must also be released after preparation, generation, validation,
-repair, or collaborator failure. One failed or canceled run must not reduce
-the backend's future usable capacity.
-
-Elapsed `Run` timing will include time spent waiting after the call is
-accepted. `PreparedRun` timing will continue to describe preparation rather
-than queue waiting.
-
-## Engine And Client Boundaries
-
-All pools and queued state belong to one `Engine`. Separate engines do not
-share capacity, even when they register the same backend ID or endpoint. The
-feature introduces no process-global scheduler.
-
-The engine will apply policy consistently to the built-in model client and an
-injected `LLMClient`. Consumers calling their own client outside Promptkit are
-outside this boundary. Injected clients remain responsible for their internal
-thread safety and cancellation behavior.
-
-Backend policy is keyed by the resolved backend ID rather than endpoint text.
-This preserves stable routing when a selected backend's endpoint is overridden
-and avoids accidentally combining unrelated registrations that happen to use
-the same URL.
-
-## Queue Lifetime And Observability
-
-Admission state is buffered, ephemeral, and in-process. It is not persisted
-and has no survival guarantee across engine disposal or process termination.
-Promptkit will not introduce background job ownership or require consumers to
-start or stop workers.
-
-The initial feature does not require public queue-depth metrics, callbacks, or
-inspection APIs. Capacity errors and ordinary call timing provide the
-consumer-visible behavior. Operational observability can be added later
-without coupling the scheduling mechanism to an application logging or
-metrics system.
-
-## Compatibility
-
-Consumer-registered backends and endpoint-only profiles remain unlimited
-unless concurrency is explicitly configured, preserving their existing
-behavior.
-
-The built-in OpenRouter backend will change from unlimited concurrency to a
-limit of 16 with a bounded waiting capacity of 1024. Ordinary synchronous
-calls remain unchanged, while unusually high concurrent use may now wait or
-return the capacity error. This behavioral change must be identified in the
-release notes for the version that publishes it.
-
-Adding backend policy fields and a public capacity error is otherwise
-additive. The change will use a pre-`v1` minor release under Promptkit's
-[release policy](../release.md#release-model).
-
-## Non-Goals
-
-This scope does not include:
-
-- asynchronous job handles, polling, or detached result delivery;
-- durable or cross-process queues;
-- persistence or recovery across engine or process shutdown;
-- priorities, scheduling weights, or consumer-defined fairness classes;
-- automatic retries, backoff, rate-limit interpretation, or provider quota
-  discovery;
-- token-per-minute or request-per-minute rate limiting;
-- dynamic reconfiguration after engine construction;
-- per-profile or per-run concurrency overrides;
-- endpoint-keyed pooling for profiles without a backend ID;
-- process-global coordination across engines;
-- application worker lifecycle, logging, tracing, or metrics policy; or
-- changes to prompt, profile, schema, or model-provider wire formats.
-
-## Target End State
-
-This roadmap reaches its target end state when:
-
-- each engine independently coordinates configured backend capacity;
-- the built-in OpenRouter backend allows 16 active generations and up to 1024
-  waiting runs;
-- consumer backends can opt into their own active and waiting limits while
-  remaining unlimited by default;
-- endpoint overrides retain the selected backend's capacity pool and
-  endpoint-only profiles remain unlimited;
-- synchronous `Run` callers wait for and receive their ordinary result;
-- admission is bounded before expensive preparation work where practical;
-- every initial and repair generation observes the backend's active limit
-  without serializing preparation or validation;
-- a full waiting queue returns a recognizable capacity error without invoking
-  the model client;
-- cancellation and all failure paths promptly release capacity and preserve
-  context error identity;
-- built-in and injected model clients receive the same scheduling behavior;
-- pools remain ephemeral, engine-scoped, and independent across backend IDs;
-  and
-- current-state GoDoc, consumer, internal, and release documentation describe
-  the implemented behavior once it lands.

docs/roadmap/deferred.md

@@ -0,0 +1,82 @@
+# Deferred Feature Ideas
+
+## Purpose
+
+This document catalogs feature ideas that remain potentially useful but have
+been deliberately postponed. These ideas are not awaiting ordinary selection
+from the [future feature catalog](future.md); each has a stated reason to wait
+and should be reconsidered only when its trigger becomes relevant.
+
+Deferred entries are not commitments, schedules, active implementation plans,
+or descriptions of current behavior. When an entry is reactivated, move it to
+`future.md` for evaluation or directly into a focused roadmap after its open
+design dependencies have been resolved.
+
+## Deferred Ideas
+
+### Semantic Execution-Target Fingerprints
+
+**Reason for deferral:** A stable digest requires a deliberate semantic-
+equality and versioning design. Notarius can safely use conservative source
+hashes and a Promptkit release marker today, while Weatherreporter does not
+currently reuse LLM-dependent checkpoints.
+
+Promptkit could expose an opaque equality value for a resolved profile and its
+effective generation target. This would let checkpointing consumers detect
+generation-affecting configuration changes without hashing YAML presentation
+or depending on Promptkit's built-in catalog layout.
+
+The digest should change with semantically relevant state such as the resolved
+model, endpoint, backend routing identity, request defaults, extra parameters,
+profile generation settings, and selected built-in profile semantics. It
+should exclude credential values, concurrency and queue policy, source paths,
+comments, formatting, and other representation-only changes. Whether a
+credential environment-variable name affects equality must be decided
+explicitly. The encoding should remain opaque and internally versioned so
+Promptkit can deliberately invalidate earlier digests when its resolution
+semantics change.
+
+Reconsider this idea when a downstream consumer needs Promptkit-owned
+checkpoint equality or when a broader semantic identity design is selected.
+
+### Eager Source Validation
+
+**Reason for deferral:** Exact prompt and profile inspection may already
+provide a sufficiently small validation surface. Experience from downstream
+adoption should establish whether an engine-wide operation would add enough
+value to justify its broader contract.
+
+Promptkit could provide an explicit offline operation that discovers and
+structurally validates configured prompt, profile, and schema sources without
+model generation. The normal `NewEngine` path would remain lazy.
+
+An eager operation would need coherent handling for duplicate prompt IDs and
+versions, strict YAML decoding, referenced content files, profile/backend
+membership, schema syntax and transitive references, context cancellation,
+and source-specific public errors. Credential declarations must remain
+separate from credential values; checking current environment availability,
+if supported at all, should be an explicit option and must not expose secrets.
+
+Reconsider this idea after downstream use of `InspectPrompt`,
+`InspectProfile`, and fixture-based preparation demonstrates a concrete gap.
+
+### Structured Generation Errors
+
+**Reason for deferral:** Existing `ErrLLMGenerate` classification, preserved
+injected-client errors, and prepared execution details currently provide the
+necessary failure boundary. A typed error should wait for stronger downstream
+demand and a transport-neutral field design.
+
+Promptkit could expose safe structured generation context through
+`errors.As` while preserving `errors.Is(err, ErrLLMGenerate)`. Potential
+fields include the selected backend ID and model plus an optional HTTP status
+when the built-in OpenAI-compatible transport supplies one.
+
+The design must not expose provider response bodies, endpoints, credential
+environment names or values, request content, or generated content. It should
+not duplicate prompt and profile provenance already available from a prepared
+execution, and it must preserve the identity of errors returned by injected
+clients. Retry and backoff policy remains a consumer responsibility.
+
+Reconsider this idea when consumers need structured generation diagnostics
+beyond the existing sentinel, wrapped client error, and preparation record.

docs/roadmap/future.md

@@ -12,6 +12,9 @@ consumer value, and important scope boundaries. Defer API design,
 implementation details, sequencing, and acceptance criteria until an idea is
 selected.
 
+Ideas that have been deliberately postponed rather than left available for
+ordinary selection belong in the [deferred catalog](deferred.md).
+
 ## Using This Catalog
 
 - Add an idea when its purpose and likely value can be stated clearly.
@@ -23,6 +26,8 @@ selected.
 - When an idea is selected, move its active planning to a focused roadmap or,
   when it requires a durable architectural decision, an ADR. Update
   current-state documentation only when implementation lands.
+- Move an idea to `deferred.md` when maintainers decide to retain it but wait
+  for a stated design dependency, demand signal, or reconsideration trigger.
 - Remove ideas that are no longer relevant. Retain a rejected idea only when
   its rationale is likely to prevent repeated reconsideration.
 
@@ -33,9 +38,7 @@ consumers.
 
 ## Ideas
 
-No ideas are currently cataloged. Backend-specific concurrency management has
-been selected for active planning in the
-[focused concurrency roadmap](concurrency.md).
+No ideas currently await selection.
 
 ## Entry Format
 

docs/roadmap/implementation.md

@@ -1,841 +0,0 @@
-# Backend-Specific Concurrency Management Implementation Plan
-
-**Status:** Complete.
-
-## Purpose
-
-This document is the decision-complete implementation plan for
-[backend-specific concurrency management](concurrency.md). It is written for a
-coding agent that will implement each stage in order.
-
-The feature roadmap owns the intended capability, consumer value, policy
-choices, compatibility decision, and target end state. This document owns the
-concrete API, internal representation, scheduling architecture, implementation
-sequence, test ownership, documentation updates, and completion gates.
-
-## Implementation Rules
-
-- Complete the stages in order. Keep the repository compiling and the focused
-  tests passing at every stage boundary.
-- Preserve unrelated working-tree changes. In particular, `concurrency.md` and
-  the removal of its source idea from `future.md` may already be uncommitted
-  when implementation begins; retain both.
-- Follow every policy under `docs/policy/`, the task-specific reading guide in
-  `docs/development.md`, and the target behavior in `concurrency.md`.
-- Keep the public API in the root `promptkit` package and implementation
-  details under `internal/`. Do not expose scheduler types or create another
-  public package.
-- Use only the Go standard library for scheduling. Do not add a queue,
-  semaphore, worker-pool, or metrics dependency.
-- Preserve synchronous, wait-for-result `Run`, unrestricted `Prepare`,
-  engine-local state, endpoint-only profiles, backend-selected profiles,
-  backend identity through endpoint overrides, and injected `LLMClient`
-  behavior.
-- Do not broaden the work into asynchronous jobs, durable queues, retries,
-  rate limiting, dynamic configuration, priorities, worker lifecycle,
-  endpoint-keyed pools, or public queue observability.
-- Keep all tests deterministic, bounded, offline, and race-safe. Coordinate
-  concurrent tests with channels and barriers rather than timing assumptions
-  or live providers.
-- Update exact GoDoc with each exported declaration change. Update durable
-  current-state documents only after the corresponding behavior is
-  implemented.
-- Test configurable mechanisms with small test-owned limits. Assert the exact
-  OpenRouter `16` and default queue `1024` values only at the registry contract
-  that owns those operational defaults.
-- Do not create a release, change a module version, or tag a commit. The final
-  implementation handoff must identify the built-in OpenRouter behavior change
-  for the next pre-`v1` minor release.
-
-## Fixed Design
-
-### Public Backend Configuration
-
-Append these fields to the existing root `Backend` type in `backends.go`:
-
-```go
-type Backend struct {
-	// Existing fields remain unchanged and in their current order.
-
-	ConcurrencyLimit int
-	QueueCapacity    *int
-}
-```
-
-Use these exact semantics:
-
-| Public values | Meaning |
-| --- | --- |
-| `ConcurrencyLimit == 0`, `QueueCapacity == nil` | Unlimited backend; preserve current behavior. |
-| `ConcurrencyLimit > 0`, `QueueCapacity == nil` | Limit active generations and use the default waiting capacity of 1024. |
-| `ConcurrencyLimit > 0`, `QueueCapacity != nil` | Limit active generations and use the pointed-to capacity exactly, including zero. |
-| `ConcurrencyLimit < 0` | Invalid engine configuration. |
-| `QueueCapacity != nil` and `*QueueCapacity < 0` | Invalid engine configuration. |
-| `ConcurrencyLimit == 0` and `QueueCapacity != nil` | Invalid engine configuration because a queue without an active limit has no defined consumer value. |
-
-`ConcurrencyLimit` counts simultaneous calls to the engine-owned internal
-model-client boundary for this backend. `QueueCapacity` controls additional
-accepted `Run` invocations beyond that limit. The maximum admitted runs for a
-limited backend is therefore:
-
-```text
-ConcurrencyLimit + effective QueueCapacity
-```
-
-Guard that addition against integer overflow during backend validation.
-Do not impose an arbitrary upper bound beyond non-negativity and overflow
-safety.
-
-The `QueueCapacity` pointer exists only to distinguish omission from explicit
-zero. `WithBackend` and `NewEngine` must not retain the caller's pointer.
-`Backend` continues to have no stable JSON representation, and consumers
-remain directed to keyed literals.
-
-Do not add concurrency fields to `Profile`, `ExecutionTarget`,
-`ExecutionTargetOverride`, `RunRequest`, prompt or profile files, or stable
-prepared/result JSON.
-
-### Built-In And Custom Defaults
-
-The backend registry owns these exact operational defaults:
-
-```go
-const (
-	openRouterConcurrencyLimit = 16
-	defaultQueueCapacity       = 1024
-)
-```
-
-The built-in `openrouter` definition has a normalized concurrency limit of 16
-and queue capacity of 1024.
-
-Consumer registrations remain unlimited when concurrency is omitted. For a
-consumer backend with a positive limit and omitted queue capacity, normalize
-the queue capacity to 1024. Preserve an explicitly configured zero.
-
-Consumers still cannot replace the reserved `openrouter` registration.
-Endpoint-only profiles have no backend policy and remain unlimited. A selected
-backend retains its pool when a profile or request overrides only its endpoint.
-
-### Internal Backend Representation
-
-Extend `internal/domain.Backend` with scalar policy values and explicit
-presence rather than retaining a pointer:
-
-```go
-type Backend struct {
-	// Existing fields...
-	ConcurrencyLimit int
-	QueueCapacity    int
-	QueueCapacitySet bool
-}
-
-type BackendCapacityPolicy struct {
-	ConcurrencyLimit int
-	QueueCapacity    int
-}
-```
-
-`WithBackend` converts the public pointer into `QueueCapacity` plus
-`QueueCapacitySet`. Registry normalization validates the combinations above,
-fills the default, and leaves every limited stored backend with
-`QueueCapacitySet == true`. Unlimited stored backends retain zero values and
-`QueueCapacitySet == false`.
-
-Add this internal registry method:
-
-```go
-func (r *Registry) CapacityPolicies() map[string]domain.BackendCapacityPolicy
-```
-
-It returns a newly allocated map containing only limited backends. Values are
-scalars, so callers cannot mutate registry state. The built-in OpenRouter
-policy is included. `GetBackend` continues returning a defensive backend copy,
-now including normalized scalar capacity metadata.
-
-Capacity policy is operational registry metadata. Do not merge it into an
-execution target or expose it to injected model clients.
-
-### Public Capacity Error
-
-Add this root sentinel beside the other run errors in `engine.go`:
-
-```go
-var ErrCapacityExceeded = errors.New("backend capacity exceeded")
-```
-
-Its GoDoc must state that it identifies a `Run` rejected because the selected
-backend has already admitted `ConcurrencyLimit + QueueCapacity` runs. It is
-not an invalid request, an LLM/provider rate-limit response, or an
-`ErrLLMGenerate` failure.
-
-The internal capacity component owns a corresponding internal
-`ErrCapacityExceeded`. Add its mapping in `publicErrorFor` before the broader
-generation and invalid-request cases. The public error must preserve the
-internal error through wrapping while matching `ErrCapacityExceeded` with
-`errors.Is`.
-
-A capacity rejection returns no partial result and must not invoke the
-artifact reader, renderer, schema loader, validator, or model client. Prompt,
-profile, and backend loading needed to select the pool may already have
-occurred.
-
-### Internal Capacity Component
-
-Add `internal/capacity` as the single owner of engine-local run admission and
-active-generation permits.
-
-Use these package-level boundaries:
-
-```go
-var ErrCapacityExceeded error
-
-type Manager struct {
-	// Private immutable pool map.
-}
-
-func NewManager(
-	policies map[string]domain.BackendCapacityPolicy,
-) (*Manager, error)
-
-func (m *Manager) Admit(
-	ctx context.Context,
-	backendID string,
-) (release func(), err error)
-
-func NewClient(m *Manager, next llm.Client) llm.Client
-```
-
-`NewManager` copies the supplied map and creates one independent pool per
-limited backend. Defensively reject blank IDs, non-positive concurrency
-limits, negative queue capacities, or total-capacity overflow even though the
-registry normally supplies normalized values. Construction creates no worker
-goroutines.
-
-An absent manager, blank backend ID, or ID absent from the policy map is
-unlimited:
-
-- `Admit` succeeds with a non-nil no-op release function; and
-- the client wrapper calls the next client directly.
-
-For a limited pool, `Admit` is immediate and context-aware:
-
-1. return `ctx.Err()` if the context is already done;
-2. under the pool lock, compare admitted runs with
-   `ConcurrencyLimit + QueueCapacity`;
-3. return an error matching internal `ErrCapacityExceeded` when full; or
-4. increment admitted runs and return an idempotent release function.
-
-The release function decrements admission exactly once, even if accidentally
-called more than once. It does not release an active-generation permit; those
-permits have their own lifetime.
-
-### FIFO Generation Permits
-
-`NewClient` returns an internal `llm.Client` wrapper around either the built-in
-client or the public-client adapter. It must preserve requests, successful
-responses, nil responses, and collaborator error identities exactly.
-`next` must be non-nil; `NewEngine` and internal runner construction maintain
-that invariant. A nil manager returns `next` unchanged.
-
-For a configured backend ID, the wrapper:
-
-1. acquires one active-generation permit from the matching pool;
-2. waits in FIFO order when the active count equals `ConcurrencyLimit`;
-3. removes a canceled waiter and returns `ctx.Err()` when cancellation wins
-   before the permit is granted;
-4. invokes the next client only after a permit is granted; and
-5. releases the permit with `defer` after every success, nil response,
-   collaborator error, panic unwinding, or context outcome.
-
-Implement FIFO and cancellation explicitly with a mutex and an ordered waiter
-list. A channel used only as a counting semaphore is insufficient because it
-does not define FIFO ordering or safe removal of canceled waiters.
-
-Permit grant and cancellation must have one lock-protected linearization
-point. If cancellation removes the waiter first, do not invoke the next
-client. If grant wins first, invoke the next client with the caller's context;
-the next client may then observe cancellation normally. Never lose or
-double-release a permit in this race.
-
-Releasing a permit transfers it to the oldest non-canceled waiter before
-making it generally available. Different backend pools never share admission
-or active counts.
-
-The active wrapper enforces its limit even if an internal caller invokes it
-without a run admission lease. Bounded backlog is guaranteed for ordinary
-engine `Run` calls by the runner admission path; no public API exposes the
-wrapped internal client directly.
-
-### Engine Assembly
-
-In `NewEngine`, after constructing the validated backend registry:
-
-1. obtain `backendRegistry.CapacityPolicies()`;
-2. construct one `capacity.Manager`;
-3. construct the selected base internal LLM client exactly as today;
-4. wrap that base client with `capacity.NewClient`; and
-5. pass both the wrapped client and manager-as-admitter to the runner.
-
-Every `NewEngine` call constructs a distinct manager. Do not cache managers,
-pools, or policies in package globals. The wrapper must be applied after a
-public injected client is adapted to `internal/llm.Client`, so built-in and
-injected clients receive identical scheduling behavior.
-
-If `NewManager` reports a defensive configuration error, make `NewEngine`
-return an error matching `ErrInvalidConfig`.
-
-`Prepare` does not use the manager. An injected client remains required to be
-safe for concurrent calls because different backend pools and unlimited
-backends may still invoke it concurrently.
-
-### Shared Two-Phase Preparation
-
-Refactor `internal/usecase.Runner` so `Prepare` and `Run` share one preparation
-pipeline with two private phases. Do not duplicate prompt/profile/backend
-selection or execution precedence.
-
-The first phase resolves only the state required before admission:
-
-1. validate `PromptID`;
-2. normalize the direct session ID;
-3. load the prompt definition;
-4. hash the original prompt definition at its existing error-order position;
-5. select and load the execution profile;
-6. resolve the selected backend;
-7. resolve and validate the effective execution target and credentials; and
-8. resolve the effective output contract without loading its schema.
-
-Return a private state value containing the loaded definition, normalized
-direct session, prompt-definition hash, selected profile ID, effective target,
-numeric-presence metadata, effective output contract, and preparation start
-time. Keep this value private to `internal/usecase`.
-
-The second phase consumes that state and performs:
-
-1. structured-output schema loading;
-2. artifact loading and input hashing;
-3. message and prompt-session rendering;
-4. direct-session application;
-5. rendered-prompt hashing; and
-6. `PreparedRun` construction and timing.
-
-Preserve every existing precedence rule, error identity, direct-session
-template bypass, hash input, selected identity, copy guarantee, and timing
-field. Do not reload the prompt, profile, or backend between phases.
-
-`Runner.Prepare` records its start time, runs both phases consecutively, and
-never calls admission. Its behavior and error ordering remain unchanged.
-
-`Runner.Run` records its existing run start time, runs the first preparation
-phase, and then calls:
-
-```go
-release, err := r.admitter.Admit(ctx, effectiveBackendID)
-```
-
-Use a narrow use-case-owned interface with the same signature:
-
-```go
-type RunAdmitter interface {
-	Admit(context.Context, string) (func(), error)
-}
-```
-
-A nil admitter means unlimited behavior for internal constructors and tests.
-On successful admission, immediately `defer release()` around the remainder of
-the run. Then run the second preparation phase, initial generation,
-validation, and all repair attempts.
-
-If admission returns internal `capacity.ErrCapacityExceeded`, add useful
-backend context without changing its identity. If it returns `ctx.Err()`,
-preserve that identity directly rather than recategorizing it as invalid
-request or generation failure.
-
-This refactor intentionally replaces the current literal `Run`-calls-`Prepare`
-implementation with shared private phases. Update current-state documentation
-to describe one shared pipeline rather than retaining an inaccurate call-graph
-claim.
-
-### Generation And Repair Lifetime
-
-The admission lease covers the entire accepted run:
-
-- second-phase preparation;
-- initial generation;
-- validation;
-- every repair; and
-- all failure and cancellation exits.
-
-Preparation and validation do not hold an active-generation permit. The
-wrapped client acquires a permit only around each actual `Generate` call.
-
-The runner's initial generation already carries the effective backend ID in
-`GenerateRequest.Target`. Preserve that value. `RepairRequest.Target` and the
-default repairer's generated request must continue carrying the same backend
-ID, allowing each repair to reacquire the same pool's active permit.
-
-When testing or constructing `NewRunnerWithRepairer`, pass the same wrapped
-client to both the runner and `NewDefaultOutputRepairer`. Do not add capacity
-state to `RepairRequest`, `ExecutionTarget`, or public generation values.
-
-A repair remains within its existing admission lease. It waits for a FIFO
-active permit but never performs a second bounded admission and therefore
-cannot fail merely because later runs filled the admission capacity.
-
-### Error And Cancellation Semantics
-
-The required public outcomes are:
-
-| Situation | Required error identity |
-| --- | --- |
-| Admission capacity is full | `ErrCapacityExceeded` only; not `ErrInvalidRequest` or `ErrLLMGenerate`. |
-| Context is done before admission succeeds | Preserve `ctx.Err()`; do not return capacity exhaustion. |
-| Context cancels while waiting for an active permit | Preserve `ctx.Err()` through the existing `ErrLLMGenerate` generation category. |
-| Wrapped client fails after permit acquisition | Preserve existing `ErrLLMGenerate` and collaborator identities. |
-| Preparation or validation fails after admission | Preserve its existing category and release admission. |
-
-Maintain the existing rule that `Run` returns no partial result on any
-operational error. Do not add queue status to errors or results.
-
-`RunResult.Duration` continues to start at runner entry and therefore includes
-pre-admission resolution, accepted preparation, and active-permit waiting.
-`PreparedRun.DurationMS` continues to cover only its shared preparation phases;
-it does not include later generation waiting. Capacity-rejected calls have no
-result or timing value.
-
-### Ownership And Concurrency Safety
-
-The registry, capacity policy map, pool map, and per-pool limits are immutable
-after engine construction. Only admission counts, active counts, and waiter
-lists are mutable and must be protected by the owning pool mutex.
-
-Do not retain public queue pointers, caller request values, contexts, or
-generation requests after their call completes. A canceled waiter must be
-unlinked so its context and request cannot remain reachable from the pool.
-
-Do not hold a pool mutex while:
-
-- loading or rendering prompts;
-- reading artifacts or schemas;
-- invoking a model client;
-- validating output;
-- closing a waiter notification channel if the implementation could re-enter
-  pool code; or
-- calling consumer code.
-
-No scheduler operation may spawn a goroutine whose lifetime outlasts the
-calling `Run`. The zero steady-state goroutine count is part of the
-in-process/no-worker-lifecycle design.
-
-## Test Ownership
-
-Use this ownership split and avoid repeating the full policy matrix at every
-layer:
-
-- `internal/backend/registry_test.go` owns normalization, validation, the exact
-  OpenRouter policy, the custom default queue, explicit zero, unlimited
-  omission, and policy-map copying.
-- `internal/capacity/manager_test.go` owns admission bounds, idempotent release,
-  FIFO active permits, cancellation races, capacity recovery, independent
-  pools, unlimited IDs, and observed peak concurrency.
-- `internal/capacity/client_test.go` owns wrapper request/response/error
-  transparency and the rule that cancellation before grant does not invoke the
-  next client. Combine these with manager tests if one coherent package test
-  expresses the behavior more clearly.
-- `internal/usecase/runner_test.go` owns two-phase preparation parity, pool
-  selection, admission before expensive work, admission release across run
-  exits, `Prepare` bypass, and repair reuse of the admitted backend.
-- Root external-package tests own public configuration conversion, assembled
-  engine-local behavior, endpoint-override routing, injected-client limiting,
-  and public capacity/context error identities.
-- Existing model-client HTTP tests remain unchanged because scheduling does
-  not alter the OpenAI-compatible wire contract.
-
-Concurrency tests must use test-owned limits such as one or two and
-channel-controlled blocking clients. Record observed active and peak counts
-under a mutex or atomics. Do not use `time.Sleep` to infer queue state.
-Package-internal tests may inspect a waiter list under its mutex through a
-small test helper when necessary to establish deterministic FIFO ordering; do
-not add production metrics or hooks solely for tests.
-
-Do not add separate tests for trivial scalar copies when registry or assembled
-behavior already protects them.
-
-## Stage 1 — Backend Policy And Public Configuration
-
-**Status:** Complete.
-
-### Goal
-
-Add the public and internal backend policy representation, normalize all
-configured states, and expose immutable normalized policies without changing
-runtime scheduling yet.
-
-### Work
-
-1. Add `ConcurrencyLimit` and `QueueCapacity` to `Backend` in `backends.go`
-   with exact GoDoc for unlimited, defaulted, explicit-zero, invalid, and
-   engine-scoped behavior.
-2. Convert the public queue pointer into scalar value plus presence in
-   `WithBackend`; do not retain the pointer.
-3. Add the internal backend policy fields and
-   `BackendCapacityPolicy` to `internal/domain/domain.go`.
-4. Add the two registry-owned constants and configure the built-in OpenRouter
-   definition with 16 and 1024.
-5. Extend `normalizeBackend` with the fixed validation, defaulting, explicit
-   zero, and overflow rules.
-6. Add `Registry.CapacityPolicies`, returning only limited policies in a fresh
-   map.
-7. Update existing backend composite literals and assertions only where the
-   new fields are relevant. Continue using keyed literals.
-
-### Tests
-
-1. Extend the exact built-in registry test with the OpenRouter limit and queue.
-2. Add one coherent table covering unlimited omission, default queue,
-   explicit-zero queue, negative values, queue-without-limit, and total
-   overflow.
-3. Extend the registry copy/normalization test to prove returned policy maps
-   cannot mutate registry state.
-4. Add root coverage only if needed to prove the public pointer/presence
-   conversion; do not reproduce registry validation cases at the facade.
-
-### Focused Validation
-
-Run:
-
-```sh
-gofmt -w backends.go internal/domain/domain.go \
-  internal/backend/registry.go internal/backend/registry_test.go
-go test . ./internal/backend
-go vet . ./internal/backend
-git diff --check
-```
-
-Include another touched Go test file in `gofmt` only if it actually changed.
-
-### Completion Gate
-
-This stage is complete when every public configuration state has one normalized
-internal meaning, OpenRouter exposes exactly 16/1024, custom backends remain
-unlimited by omission, and no runtime call is scheduled yet.
-
-## Stage 2 — Engine-Local Capacity Manager
-
-**Status:** Complete.
-
-### Goal
-
-Implement and prove the bounded admission mechanism and FIFO active-generation
-client wrapper independently of runner orchestration.
-
-### Work
-
-1. Add `internal/capacity/manager.go` with the manager, immutable policy copy,
-   per-backend pools, internal error, immediate admission, idempotent release,
-   and FIFO context-aware active permits.
-2. Add `internal/capacity/client.go` with the transparent `llm.Client` wrapper.
-3. Use mutex-protected waiter state and an ordered list; explicitly resolve
-   grant-versus-cancel races.
-4. Ensure unlimited and independent-pool fast paths avoid queue allocation.
-5. Do not start workers, timers, cleanup goroutines, or process-global state.
-
-### Tests
-
-1. Add a compact constructor-validation table for blank IDs, non-positive
-   limits, negative queues, and total-capacity overflow.
-2. With a small configured policy, prove that exactly
-   `limit + queueCapacity` admissions succeed, the next matches
-   `ErrCapacityExceeded`, and a release permits another admission.
-3. Prove release is idempotent.
-4. Drive more blocked client calls than the active limit and assert observed
-   peak concurrency never exceeds that limit.
-5. Prove FIFO order with deterministic queue-entry synchronization.
-6. Cancel the first and a middle waiter and prove they are removed, never call
-   the wrapped client, and do not block later waiters.
-7. Exercise the grant/cancel race repeatedly under `go test -race`, asserting
-   no permit leak or double invocation.
-8. Prove different backend IDs proceed independently and blank, unknown, or
-   nil-manager paths remain unlimited.
-9. Prove request values, successful and nil responses, and collaborator errors
-   pass through unchanged after permit acquisition.
-
-### Focused Validation
-
-Run:
-
-```sh
-gofmt -w internal/capacity/manager.go \
-  internal/capacity/manager_test.go \
-  internal/capacity/client.go \
-  internal/capacity/client_test.go
-go test ./internal/capacity
-go test -race ./internal/capacity
-go vet ./internal/capacity
-git diff --check
-```
-
-If tests are combined into one file, omit the nonexistent file from `gofmt`.
-
-### Completion Gate
-
-This stage is complete when the standalone component enforces relational
-admission and active limits, FIFO cancellation is race-safe, separate pools
-are independent, and the wrapper is transparent apart from waiting.
-
-## Stage 3 — Shared Preparation And Early Run Admission
-
-**Status:** Complete.
-
-### Goal
-
-Refactor runner preparation into one shared two-phase pipeline and place
-bounded admission after backend resolution but before schema, artifact, and
-rendering work.
-
-### Work
-
-1. Add the private pre-admission preparation state and split the existing
-   `Prepare` logic according to the fixed design.
-2. Make `Runner.Prepare` call both phases without an admitter.
-3. Add the `RunAdmitter` interface and runner field.
-4. Update `NewRunner` and `NewRunnerWithRepairer` to accept the optional
-   admitter; update internal call sites with nil until root assembly is wired.
-5. Change `Runner.Run` to use the first phase, admit by effective backend ID,
-   defer the returned release, and then use the second phase.
-6. Preserve all existing error precedence, target resolution, hashes,
-   metadata, session behavior, and timing.
-7. Return capacity and context errors with the fixed identities. Do not invoke
-   later collaborators after rejection.
-
-### Tests
-
-1. Keep the existing `Run`/`Prepare` parity coverage passing to prove the
-   shared phases do not drift.
-2. Add a fake admitter that records backend IDs and release calls.
-3. Prove a backend-selected run admits with the selected ID even when the
-   endpoint is overridden.
-4. Prove an endpoint-only run uses the unlimited/blank identity and that
-   `Prepare` never calls admission.
-5. Reject admission and assert schema, artifact, renderer, validator, repairer,
-   and LLM collaborators are not invoked.
-6. Prove admission is released after one successful run and representative
-   second-phase, generation, and validation errors. Prefer a small table around
-   the single `defer` invariant rather than duplicating every error test.
-7. Retain direct-session, backend precedence, credential, hashing, and repair
-   tests unchanged except for constructor arguments.
-
-### Focused Validation
-
-Run:
-
-```sh
-gofmt -w internal/usecase/runner.go \
-  internal/usecase/runner_test.go
-go test ./internal/usecase
-go test -race ./internal/usecase
-go vet ./internal/usecase
-git diff --check
-```
-
-### Completion Gate
-
-This stage is complete when `Prepare` remains unrestricted, `Run` admits after
-one canonical routing phase and before expensive completion work, every exit
-releases admission, and existing preparation semantics remain unchanged.
-
-## Stage 4 — Engine Assembly And Public Runtime Contract
-
-**Status:** Complete.
-
-### Goal
-
-Wire one manager into each engine, schedule built-in and injected clients,
-expose the capacity error, and prove assembled runtime behavior.
-
-### Work
-
-1. Add public `ErrCapacityExceeded` and its exact GoDoc in `engine.go`.
-2. Map internal capacity exhaustion in `errors.go`.
-3. Construct the manager from the registry policy snapshot in `NewEngine`.
-4. Wrap the selected internal client after built-in or injected-client
-   selection and pass the manager and wrapped client to the runner.
-5. Update `Engine`, `NewEngine`, `Run`, `WithLLMClient`, and `LLMClient` GoDoc
-   only where concurrency, capacity, or cancellation statements change.
-6. Ensure manager-construction errors match `ErrInvalidConfig`.
-7. For internal repair coverage, construct the default repairer with the same
-   wrapped client used by its runner and confirm repair target backend identity
-   remains intact.
-
-### Tests
-
-1. Add an external-package assembled test with a small custom limit and a
-   blocking injected client; assert peak generation equals or remains below
-   the configured limit.
-2. With queue capacity zero, block one accepted run before generation and
-   assert the next matching-backend run returns `ErrCapacityExceeded`, does not
-   match `ErrInvalidRequest` or `ErrLLMGenerate`, returns no result, and never
-   reaches expensive collaborators or the client.
-3. In the same or another focused workflow, prove an endpoint override remains
-   in the selected backend's pool.
-4. Prove two engines with the same backend ID have independent capacity.
-5. Prove an unlimited custom backend and an endpoint-only profile preserve
-   concurrent behavior.
-6. Cancel a call waiting for an active permit; assert it matches both
-   `context.Canceled` and `ErrLLMGenerate`, never invokes the injected client,
-   and leaves capacity reusable.
-7. Add one internal repair workflow with concurrent runs or controlled permits
-   showing initial and repair generations never exceed the same backend limit
-   and repairs do not perform a second admission.
-8. Extend the public error sentinel contract test with
-   `ErrCapacityExceeded`.
-
-Avoid a second HTTP-level concurrency suite: the capacity client tests and one
-assembled injected-client workflow already protect the shared wrapper used by
-the built-in client.
-
-### Focused Validation
-
-Run:
-
-```sh
-gofmt -w engine.go errors.go backends.go \
-  internal/usecase/runner.go internal/usecase/runner_test.go \
-  engine_test.go public_contract_test.go
-go test . ./internal/backend ./internal/capacity ./internal/usecase
-go test -race . ./internal/capacity ./internal/usecase
-go vet . ./internal/backend ./internal/capacity ./internal/usecase
-git diff --check
-```
-
-Add any newly created capacity files to `gofmt` when they changed in this
-stage.
-
-### Completion Gate
-
-This stage is complete when every engine has independent pools, limited runs
-are bounded and FIFO at generation, endpoint routing is correct, capacity and
-context errors are stable, repairs reuse admission, and both client kinds pass
-through the same wrapper.
-
-## Stage 5 — Durable Documentation And Final Validation
-
-**Status:** Complete.
-
-### Goal
-
-Move implemented contracts into their durable owners, record compatibility
-impact, and validate the complete repository.
-
-### Work
-
-1. Review every changed exported declaration. Ensure GoDoc is the canonical
-   owner of exact field types, nil/zero semantics, defaulting, error identity,
-   engine scope, concurrency safety, cancellation, and source compatibility.
-2. Update `doc.go` so its concurrency summary acknowledges backend scheduling
-   while continuing to require injected collaborators to be concurrency-safe.
-3. Update `docs/consumers/pkg-promptkit.md` with task-oriented examples for:
-   - a limited local backend;
-   - omitted queue capacity selecting 1024;
-   - explicit zero queue capacity; and
-   - handling `ErrCapacityExceeded`.
-   Keep exact field semantics in GoDoc rather than duplicating a full table.
-4. Add `docs/internal/capacity.md` as the durable owner of pool lifecycle,
-   admission, FIFO active permits, cancellation, client wrapping, and test
-   ownership.
-5. Add `internal/capacity` to `docs/internal/overview.md`.
-6. Update `docs/policy/architecture.md` to include the implemented component
-   and root assembly dependency without turning policy into an API reference.
-7. Update `docs/internal/runner.md` to describe the shared two-phase
-   preparation pipeline, early bounded admission, lease lifetime, generation
-   permits, repairs, capacity failures, and cancellation.
-8. Review `docs/formats.md`; add only a concise link or clarification if needed
-   to explain that endpoint overrides preserve backend capacity identity.
-   Do not add concurrency fields to YAML.
-9. Do not change the OpenAI-compatible integration contract or
-   `docs/internal/llm.md` unless implementation changes their current
-   statements; scheduling is outside the provider wire contract and concrete
-   model-client implementation.
-10. Record in the implementation handoff that built-in OpenRouter now limits
-    active generations to 16 with queue capacity 1024 and that the release
-    must be a pre-`v1` minor release. Do not edit the release procedure or
-    create a tag.
-11. After every check passes, set `concurrency.md`, this implementation plan,
-    and each stage status to `Complete`. Do not remove the roadmaps in the
-    implementation change; lifecycle retirement follows review.
-
-### Full Validation
-
-Run the complete sequence from `docs/development.md`:
-
-```sh
-go test ./...
-go test -race ./...
-go vet ./...
-go build ./...
-go run ./examples/go-library/prepare
-gofmt -l $(git ls-files '*.go')
-git diff --check
-```
-
-The formatting command must produce no paths. Follow every added or changed
-Markdown link and confirm its target and heading exist.
-
-Also inspect:
-
-```sh
-git status --short
-git diff --stat
-git diff
-```
-
-Confirm that:
-
-- only intended backend, capacity, runner, facade, test, documentation, and
-  roadmap files changed;
-- no `go.work`, `go.work.sum`, local module replacement, credential,
-  generated binary, coverage output, or unrelated change was introduced;
-- the built-in OpenRouter policy is exactly 16/1024;
-- custom and endpoint-only backends remain unlimited by omission;
-- explicit queue zero is distinguishable from omission;
-- no capacity value enters execution targets, generated requests, stable JSON,
-  prompt/profile YAML, or provider payloads;
-- every engine owns distinct pools with no package-global mutable state;
-- every initial and repair generation uses the active permit wrapper;
-- capacity and waiter state is released on success, error, panic unwinding,
-  and cancellation;
-- concurrency tests use deterministic coordination rather than sleeps;
-- current-state documentation describes implemented behavior rather than
-  referring readers to the roadmaps; and
-- the feature and implementation roadmaps contain no unresolved work marked
-  complete.
-
-### Completion Gate
-
-The implementation is complete only when every target-end-state item in
-`concurrency.md` is implemented, race-enabled tests demonstrate the configured
-limits and cancellation safety, durable contracts no longer depend on roadmap
-prose, and the OpenRouter compatibility change is clearly reported for the
-next minor release.
-
-## Implementation Handoff
-
-Backend-specific capacity management is implemented and has passed the complete
-repository validation sequence. The built-in OpenRouter backend now permits 16
-active generations and a waiting capacity of 1024. Custom backends remain
-unlimited when their limit is omitted, and endpoint-only profiles remain
-unlimited.
-
-Publishing this behavior requires a pre-`v1` minor release. Its release notes
-must identify that unusually high concurrent OpenRouter use can now wait or
-return `ErrCapacityExceeded`. This implementation does not change a module
-version or create a tag.
-
-## Open Questions
-
-None. The feature roadmap and this plan fix the public representation,
-registry defaults, admission bound, FIFO generation behavior, early-routing
-refactor, cancellation races, error identities, engine and repair lifetimes,
-test ownership, compatibility treatment, and non-goals required for
-implementation.

engine.go

@@ -63,9 +63,10 @@ var (
 	// ErrPromptRender identifies a failure to render prompt messages or the
 	// session ID from the resolved inputs and variables.
 	ErrPromptRender = errors.New("failed to render prompt")
-	// ErrCapacityExceeded identifies a Run rejected because the selected backend
-	// already admitted ConcurrencyLimit + QueueCapacity calls. It is not an
-	// invalid request, an LLM or provider rate-limit response, or ErrLLMGenerate.
+	// ErrCapacityExceeded identifies a Run or RunPrepared rejected because the
+	// selected backend already admitted ConcurrencyLimit + QueueCapacity calls.
+	// A [CapacityError] reports the selected backend ID. It is not an invalid
+	// request, an LLM or provider rate-limit response, or ErrLLMGenerate.
 	ErrCapacityExceeded = errors.New("backend capacity exceeded")
 	// ErrLLMGenerate identifies a model-client failure or a nil successful
 	// response. Errors returned by an injected LLMClient remain available
@@ -77,12 +78,15 @@ var (
 	ErrValidation = errors.New("failed to validate output")
 )
 
-// Engine prepares and runs Promptkit prompt requests.
+// Engine inspects prompts and profiles and prepares and runs Promptkit prompt
+// requests.
 //
-// An Engine is safe for concurrent calls to [Engine.Prepare] and [Engine.Run].
-// Each Engine owns independent backend-capacity pools that coordinate Run
-// admission and model generation. Injected collaborators may still be invoked
-// concurrently across different backend pools or for unlimited backends.
+// An Engine is safe for concurrent calls to [Engine.InspectPrompt],
+// [Engine.InspectProfile], [Engine.Prepare], [Engine.PrepareExecution],
+// [Engine.Run], and [Engine.RunPrepared]. Each Engine owns independent
+// backend-capacity pools that coordinate Run and RunPrepared admission and
+// model generation. Injected collaborators may still be invoked concurrently
+// across different backend pools or for unlimited backends.
 type Engine struct {
 	runner *usecase.Runner
 }
@@ -94,9 +98,10 @@ type Config struct {
 	// It is required unless a WithPromptFS or WithPromptFile option supplies the
 	// prompt source.
 	PromptDir string
-	// ProfileDir is an optional directory whose profiles take precedence over
-	// embedded built-in profiles. An empty value selects only built-ins unless
-	// profile options are also supplied.
+	// ProfileDir is an optional ordinary configured source whose profiles take
+	// precedence over application fallback and embedded built-in profiles. An
+	// empty value selects the lower-precedence sources unless a profile-source
+	// option supplies the ordinary source.
 	ProfileDir string
 	// SchemaDir is the root for JSON Schema files. An empty value uses the
 	// current directory. WithSchemaFS or WithSchemaFile replaces this source.
@@ -115,12 +120,12 @@ type Config struct {
 // Option customizes engine construction.
 //
 // NewEngine applies options in argument order and ignores nil options. Within
-// each prompt-source, profile-source, in-memory-profile, schema-source,
-// model-client, and artifact-reader category, the last non-nil valid option
-// replaces earlier options in that category. WithBackend is the additive
-// exception: unique registrations accumulate, and a repeated backend ID is an
-// error rather than a replacement. An invalid option fails construction even
-// if a later option would replace it.
+// each prompt-source, ordinary-profile-source, fallback-profile-source,
+// in-memory-profile, schema-source, model-client, and artifact-reader
+// category, the last non-nil valid option replaces earlier options in that
+// category. WithBackend is the additive exception: unique registrations
+// accumulate, and a repeated backend ID is an error rather than a replacement.
+// An invalid option fails construction even if a later option would replace it.
 type Option interface {
 	apply(*engineOptions) error
 }
@@ -132,26 +137,30 @@ func (f optionFunc) apply(options *engineOptions) error {
 }
 
 type engineOptions struct {
-	llmClient       llm.Client
-	artifactReader  artifactadapter.Reader
-	promptDefs      promptdef.Repository
-	profiles        profile.Repository
-	memoryProfiles  profile.Repository
-	backends        []domain.Backend
-	validator       validate.Validator
-	promptSource    bool
-	profileSource   bool
-	memorySource    bool
-	validatorSource bool
-	artifactSource  bool
+	llmClient             llm.Client
+	artifactReader        artifactadapter.Reader
+	promptDefs            promptdef.Repository
+	profiles              profile.Repository
+	fallbackProfiles      profile.Repository
+	memoryProfiles        profile.Repository
+	backends              []domain.Backend
+	validator             validate.Validator
+	promptSource          bool
+	profileSource         bool
+	fallbackProfileSource bool
+	memorySource          bool
+	validatorSource       bool
+	artifactSource        bool
 }
 
-// WithLLMClient replaces the built-in model client used by [Engine.Run].
+// WithLLMClient replaces the built-in model client used by [Engine.Run] and
+// [Engine.RunPrepared].
 //
 // A nil client makes NewEngine fail with ErrInvalidConfig. The Engine schedules
 // Generate calls according to the selected backend's capacity policy, but the
 // client may still be called concurrently across different backend pools or for
-// unlimited backends. The client is not used by [Engine.Prepare].
+// unlimited backends. The client is not used by [Engine.Prepare] or
+// [Engine.PrepareExecution].
 func WithLLMClient(client LLMClient) Option {
 	return optionFunc(func(options *engineOptions) error {
 		if client == nil {
@@ -218,12 +227,12 @@ func WithPromptFile(path string) Option {
 
 // WithProfileFS loads execution profiles from fsys under root.
 //
-// Profiles from this source overlay built-in profiles. Profile YAML must use
-// api_key_env for environment-based credentials; raw API keys are rejected.
-// fsys must be non-nil and root must be non-empty; otherwise NewEngine fails
-// with ErrInvalidConfig. This option replaces Config.ProfileDir and earlier
-// file or FS profile-source options, but remains below WithProfiles in
-// precedence.
+// Profiles from this ordinary configured source take precedence over
+// application fallback and built-in profiles. Profile YAML must use api_key_env
+// for environment-based credentials; raw API keys are rejected. fsys must be
+// non-nil and root must be non-empty; otherwise NewEngine fails with
+// ErrInvalidConfig. This option replaces Config.ProfileDir and earlier file or
+// FS profile-source options, but remains below WithProfiles in precedence.
 func WithProfileFS(fsys fs.FS, root string) Option {
 	return optionFunc(func(options *engineOptions) error {
 		if fsys == nil {
@@ -240,11 +249,12 @@ func WithProfileFS(fsys fs.FS, root string) Option {
 
 // WithProfileFile loads execution profiles from the single profile file at path.
 //
-// The profile overlays built-in profiles. Profile YAML must use api_key_env for
-// environment-based credentials; raw API keys are rejected. path must name an
-// existing non-directory file when NewEngine applies the option. This option
-// replaces Config.ProfileDir and earlier file or FS profile-source options,
-// but remains below WithProfiles in precedence.
+// The profile takes precedence over application fallback and built-in profiles.
+// Profile YAML must use api_key_env for environment-based credentials; raw API
+// keys are rejected. path must name an existing non-directory file when
+// NewEngine applies the option. This option replaces Config.ProfileDir and
+// earlier file or FS profile-source options, but remains below WithProfiles in
+// precedence.
 func WithProfileFile(path string) Option {
 	return optionFunc(func(options *engineOptions) error {
 		fsys, root, err := fileSource(path)
@@ -257,8 +267,41 @@ func WithProfileFile(path string) Option {
 	})
 }
 
+// WithFallbackProfileFS supplies application-owned fallback profile
+// definitions from fsys under root.
+//
+// Profile lookup checks, in order, profiles supplied by WithProfiles; the
+// ordinary configured source selected by WithProfileFile, WithProfileFS, or
+// Config.ProfileDir; this fallback source; and Promptkit's embedded built-in
+// profiles. Each source supplies a complete profile definition; profile fields
+// are not merged between sources. Only an absent profile ID proceeds to the
+// next source. A matching read, parse, duplicate, validation, or credential
+// format failure stops resolution.
+//
+// Files use the ordinary strict profile YAML and api_key_env credential rules.
+// Loading and validation are lazy: NewEngine validates this option's arguments
+// but does not read profile files. fsys must be non-nil and root must be
+// nonblank; otherwise NewEngine returns an error matching ErrInvalidConfig.
+// Repeating this option replaces the earlier valid fallback source.
+//
+// This option controls profile-definition lookup, not provider or generation
+// failover.
+func WithFallbackProfileFS(fsys fs.FS, root string) Option {
+	return optionFunc(func(options *engineOptions) error {
+		if fsys == nil {
+			return ErrInvalidConfig
+		}
+		if strings.TrimSpace(root) == "" {
+			return ErrInvalidConfig
+		}
+		options.fallbackProfiles = profile.NewFSRepository(fsys, root)
+		options.fallbackProfileSource = true
+		return nil
+	})
+}
+
 // WithProfiles configures in-memory profiles that take precedence over
-// configured profile files and built-in profiles.
+// ordinary configured, application fallback, and built-in profiles.
 //
 // NewEngine validates and copies every profile. IDs must be unique within one
 // call. An invalid profile, duplicate ID, or unsupported ExtraParams value
@@ -344,13 +387,7 @@ func NewEngine(cfg Config, opts ...Option) (*Engine, error) {
 		promptDefs = promptdef.NewFilesystemRepository(cfg.PromptDir)
 	}
 
-	profiles := builtin.NewRepositoryWithDirectory(cfg.ProfileDir)
-	if options.profileSource {
-		profiles = builtin.NewRepositoryWithPrimary(options.profiles)
-	}
-	if options.memorySource {
-		profiles = profile.NewOverlayRepository(options.memoryProfiles, profiles)
-	}
+	profiles := newProfileRepository(cfg.ProfileDir, options)
 
 	backendRegistry, err := backend.NewRegistry(options.backends)
 	if err != nil {
@@ -403,6 +440,26 @@ func NewEngine(cfg Config, opts ...Option) (*Engine, error) {
 	}, nil
 }
 
+func newProfileRepository(profileDir string, options engineOptions) profile.Repository {
+	repository := builtin.NewRepository()
+
+	if options.fallbackProfileSource {
+		repository = profile.NewOverlayRepository(options.fallbackProfiles, repository)
+	}
+
+	if options.profileSource {
+		repository = profile.NewOverlayRepository(options.profiles, repository)
+	} else if strings.TrimSpace(profileDir) != "" {
+		repository = profile.NewOverlayRepository(profile.NewFilesystemRepository(profileDir), repository)
+	}
+
+	if options.memorySource {
+		repository = profile.NewOverlayRepository(options.memoryProfiles, repository)
+	}
+
+	return repository
+}
+
 func fileSource(name string) (fs.FS, string, error) {
 	cleanName := strings.TrimSpace(name)
 	if cleanName == "" {
@@ -423,6 +480,93 @@ func fileSource(name string) (fs.FS, string, error) {
 	return os.DirFS(dir), filepath.ToSlash(base), nil
 }
 
+// InspectPrompt resolves one explicit prompt definition without selecting a
+// profile or starting execution work.
+//
+// InspectPrompt requires a nonblank promptID. It passes nonblank promptID and
+// promptVersion values unchanged to the engine's ordinary, case-sensitive
+// prompt selection. An empty version succeeds only when that source has one
+// selected ID; a nonempty version selects one exact ID/version pair. The
+// configured prompt source is used without merging, fallback, or enumeration.
+//
+// A successful result proves that the selected definition and any referenced
+// message content files were structurally loaded. Inputs are returned in
+// definition order. DefaultProfileID is declared metadata only and is not
+// resolved. OutputContract is the normalized declared contract, with a JSON
+// Schema path when declared but without loading or compiling that schema.
+// PromptHash is the same opaque equality value as PreparedRun.PromptHash for
+// the selected definition and observed source state; its spelling, length,
+// encoding, algorithm, and security properties are not contracts.
+//
+// This method does not return prompt bodies, templates, source paths, schemas,
+// rendered messages, or execution settings. It does not resolve a profile or
+// credential, read artifacts or schemas, render, validate, admit capacity,
+// contact a provider, or generate model output. The returned PromptInspection
+// and its input slice are caller-owned. Filesystem-backed inspection is a
+// point-in-time lookup and does not freeze a definition for later execution.
+//
+// A nil Engine returns an error matching ErrInvalidConfig. A blank prompt ID
+// matches ErrInvalidRequest. An absent exact ID or version matches
+// ErrPromptNotFound and not ErrPromptLoad. Malformed, unreadable, duplicate,
+// ambiguous, referenced-content, or hashing failures match ErrPromptLoad.
+// Cancellation during lookup matches ErrPromptLoad while preserving the
+// context error. InspectPrompt returns no partial result on error.
+func (e *Engine) InspectPrompt(
+	ctx context.Context,
+	promptID string,
+	promptVersion string,
+) (*PromptInspection, error) {
+	if e == nil || e.runner == nil {
+		return nil, fmt.Errorf("%w: engine is nil", ErrInvalidConfig)
+	}
+
+	inspection, err := e.runner.InspectPrompt(ctx, promptID, promptVersion)
+	if err != nil {
+		return nil, mapPublicError(err)
+	}
+	return fromDomainPromptInspection(inspection), nil
+}
+
+// InspectProfile resolves one explicit profile without selecting a prompt or
+// starting execution work.
+//
+// InspectProfile trims surrounding whitespace from profileID and looks up the
+// resulting nonblank ID exactly and case-sensitively through the engine's
+// in-memory, ordinary configured-source, application fallback, and built-in
+// profile precedence. It applies the framework timeout baseline, selected
+// backend, and then selected profile to EffectiveModelParams without a request
+// override. BackendID is empty for an endpoint-only profile.
+//
+// APIKeyEnv in the returned target is an environment-variable name, never its
+// value. APIKeyRequired instead reports a direct credential requirement and is
+// mutually exclusive with a nonblank APIKeyEnv. InspectProfile neither derives
+// an ID from a prompt default_profile nor checks credential availability, so an
+// absent or blank named environment variable is not an error.
+//
+// The returned ProfileInspection and all nested mutable values are
+// caller-owned. Filesystem-backed inspection is a point-in-time lookup and
+// does not freeze the profile for a later execution. This method does not load
+// a prompt, render, read artifacts or schemas, admit backend capacity, contact
+// a provider, or generate model output.
+//
+// A nil Engine returns an error matching ErrInvalidConfig. A blank profile ID
+// matches ErrInvalidRequest. An absent exact ID matches ErrProfileNotFound and
+// not ErrProfileLoad. Malformed or unreadable profile data, an unknown backend,
+// or an invalid resolved target matches ErrProfileLoad. Cancellation during
+// profile loading matches ErrProfileLoad while preserving the context error.
+// InspectProfile returns no partial result on error.
+func (e *Engine) InspectProfile(ctx context.Context, profileID string) (*ProfileInspection, error) {
+	if e == nil || e.runner == nil {
+		return nil, fmt.Errorf("%w: engine is nil", ErrInvalidConfig)
+	}
+
+	inspection, err := e.runner.InspectProfile(ctx, profileID)
+	if err != nil {
+		return nil, mapPublicError(err)
+	}
+	return fromDomainProfileInspection(inspection), nil
+}
+
 // Prepare resolves and renders a prompt request without calling an LLM.
 //
 // Prepare selects the prompt and profile, resolves any selected backend and
@@ -457,6 +601,38 @@ func (e *Engine) Prepare(ctx context.Context, req RunRequest) (*PreparedRun, err
 	return fromDomainPreparedRun(prepared), nil
 }
 
+// PrepareExecution completely prepares a prompt request without calling the
+// configured LLMClient or reserving backend admission capacity.
+//
+// The returned opaque handle is bound to this Engine and permits one
+// [Engine.RunPrepared] invocation. Preparation freezes the selected sources,
+// rendered messages, effective settings, inputs, provider structured-output
+// metadata, and validation resources needed by that invocation. The handle
+// retains a direct RunRequest.APIKey only in private execution state;
+// [PreparedExecution.Details] is credential-redacted.
+//
+// The context governs preparation only. Cancellation after this method
+// returns does not invalidate the handle or propagate to RunPrepared.
+// PrepareExecution returns the same error categories as [Engine.Prepare] and
+// returns no handle on error. A nil Engine returns an error matching
+// ErrInvalidConfig.
+func (e *Engine) PrepareExecution(ctx context.Context, req RunRequest) (*PreparedExecution, error) {
+	if e == nil || e.runner == nil {
+		return nil, fmt.Errorf("%w: engine is nil", ErrInvalidConfig)
+	}
+
+	domainReq, err := toDomainRunRequest(req)
+	if err != nil {
+		return nil, fmt.Errorf("%w: %v", ErrInvalidRequest, err)
+	}
+
+	prepared, err := e.runner.PrepareExecution(ctx, domainReq)
+	if err != nil {
+		return nil, mapPublicError(err)
+	}
+	return &PreparedExecution{internal: prepared}, nil
+}
+
 // Run prepares a request, invokes the configured LLMClient, and validates the
 // generated output.
 //
@@ -467,9 +643,10 @@ func (e *Engine) Prepare(ctx context.Context, req RunRequest) (*PreparedRun, err
 // single-pass even when OutputContract.RepairAttempts is positive.
 //
 // Run can return every error category documented by [Engine.Prepare], plus
-// ErrCapacityExceeded and ErrLLMGenerate. ErrCapacityExceeded identifies
-// rejection before artifacts, schemas, rendering, or model generation because
-// the selected backend's admission capacity is full; it does not match
+// ErrCapacityExceeded and ErrLLMGenerate. An engine admission rejection is
+// discoverable as [CapacityError] and still matches ErrCapacityExceeded. It
+// occurs before artifacts, schemas, rendering, or model generation because the
+// selected backend's admission capacity is full; it does not match
 // ErrInvalidRequest or ErrLLMGenerate. Errors from injected clients remain
 // available through errors.Is. Cancellation while waiting for model-generation
 // capacity matches both ErrLLMGenerate and the context error. Cancellation
@@ -491,3 +668,42 @@ func (e *Engine) Run(ctx context.Context, req RunRequest) (*RunResult, error) {
 	}
 	return fromDomainRunResult(result), nil
 }
+
+// RunPrepared atomically claims and executes a handle created by
+// [Engine.PrepareExecution].
+//
+// A valid owning-Engine invocation consumes the handle's one attempt before
+// credential revalidation, backend admission, generation, or validation.
+// Cancellation, capacity rejection, generation failure, operational
+// validation failure, and success all leave the handle unusable. A nil,
+// zero-value, foreign-Engine, discarded, claimed, or used handle returns an
+// error matching ErrInvalidRequest; a nil Engine returns ErrInvalidConfig and
+// does not claim the handle.
+//
+// The supplied context governs this execution attempt independently of the
+// preparation context. It covers credential revalidation, admission,
+// generation, validation, and any internal repair. Result timing begins after
+// the claim and excludes preparation and consumer-held delay.
+//
+// RunPrepared can return ErrInvalidRequest, ErrAPIKeyEnvMissing,
+// ErrCapacityExceeded, ErrLLMGenerate, or ErrValidation as applicable while
+// preserving documented collaborator and context identities. An engine
+// admission rejection is discoverable as [CapacityError] and still matches
+// ErrCapacityExceeded. A completed content-validation rejection is returned
+// in RunResult, not as an operational error. An operational error returns no
+// partial RunResult.
+func (e *Engine) RunPrepared(ctx context.Context, prepared *PreparedExecution) (*RunResult, error) {
+	if e == nil || e.runner == nil {
+		return nil, fmt.Errorf("%w: engine is nil", ErrInvalidConfig)
+	}
+
+	var internal *usecase.PreparedExecution
+	if prepared != nil {
+		internal = prepared.internal
+	}
+	result, err := e.runner.RunPrepared(ctx, internal)
+	if err != nil {
+		return nil, mapPublicError(err)
+	}
+	return fromDomainRunResult(result), nil
+}

engine_test.go

@@ -360,14 +360,14 @@ func TestEngineExecutionSettingPrecedence(t *testing.T) {
 		wantPresence promptkit.ExecutionTargetPresence
 	}{
 		{
-			name:    "framework defaults fill zero-valued profile settings",
+			name:    "unspecified provider controls retain framework timeout",
 			profile: defaultsProfile,
 			want: promptkit.ExecutionTarget{
 				Endpoint:        defaultsProfile.endpoint,
 				Model:           defaultsProfile.model,
 				Temperature:     0,
 				MaxTokens:       0,
-				TopP:            1,
+				TopP:            0,
 				TimeoutSeconds:  600,
 				ServiceTier:     defaultsProfile.serviceTier,
 				ReasoningEffort: defaultsProfile.reasoningEffort,
@@ -2301,6 +2301,8 @@ func TestSourceOptionsRejectInvalidInputs(t *testing.T) {
 		{name: "profile fs nil", opt: promptkit.WithProfileFS(nil, "profiles")},
 		{name: "profile fs empty root", opt: promptkit.WithProfileFS(fstest.MapFS{}, "")},
 		{name: "profile file empty", opt: promptkit.WithProfileFile("")},
+		{name: "fallback profile fs nil", opt: promptkit.WithFallbackProfileFS(nil, "profiles")},
+		{name: "fallback profile fs empty root", opt: promptkit.WithFallbackProfileFS(fstest.MapFS{}, "")},
 		{name: "schema fs nil", opt: promptkit.WithSchemaFS(nil, "schemas")},
 		{name: "schema fs empty root", opt: promptkit.WithSchemaFS(fstest.MapFS{}, "")},
 		{name: "schema file empty", opt: promptkit.WithSchemaFile("")},

errors.go

@@ -3,6 +3,7 @@ package promptkit
 import (
 	"errors"
 	"fmt"
+	"strings"
 
 	"gitea.maximumdirect.net/eric/promptkit/internal/capacity"
 	"gitea.maximumdirect.net/eric/promptkit/internal/profile"
@@ -14,6 +15,11 @@ func mapPublicError(err error) error {
 	if err == nil {
 		return nil
 	}
+	var internalCapacityError *usecase.CapacityError
+	if errors.As(err, &internalCapacityError) && internalCapacityError != nil &&
+		strings.TrimSpace(internalCapacityError.BackendID) != "" {
+		return &CapacityError{BackendID: internalCapacityError.BackendID}
+	}
 	publicErr := publicErrorFor(err)
 	if publicErr == nil {
 		return err

errors_internal_test.go

@@ -20,3 +20,31 @@ func TestMapPublicErrorPreservesGenerationCancellation(t *testing.T) {
 		t.Fatalf("mapped error=%v, want context.Canceled", err)
 	}
 }
+
+func TestMapPublicErrorTranslatesCapacityError(t *testing.T) {
+	internalErr := &usecase.CapacityError{BackendID: "limited"}
+
+	err := mapPublicError(internalErr)
+	var publicErr *CapacityError
+	if !errors.As(err, &publicErr) || publicErr == nil {
+		t.Fatalf("mapped error=%v, want public CapacityError", err)
+	}
+	if publicErr.BackendID != "limited" {
+		t.Fatalf("mapped backend ID=%q, want limited", publicErr.BackendID)
+	}
+	if !errors.Is(err, ErrCapacityExceeded) {
+		t.Fatalf("mapped error=%v, want ErrCapacityExceeded", err)
+	}
+	if errors.Is(err, ErrInvalidRequest) || errors.Is(err, ErrLLMGenerate) {
+		t.Fatalf("mapped capacity error has an unrelated category: %v", err)
+	}
+	var leakedInternalErr *usecase.CapacityError
+	if errors.As(err, &leakedInternalErr) {
+		t.Fatalf("mapped error exposes internal CapacityError: %v", err)
+	}
+
+	internalErr.BackendID = "changed"
+	if publicErr.BackendID != "limited" {
+		t.Fatalf("mapped backend ID changed with source error: %q", publicErr.BackendID)
+	}
+}

internal/capacity/manager_test.go

@@ -124,6 +124,11 @@ func TestManagerAdmissionHonorsContextAndUnlimitedBackends(t *testing.T) {
 	if err != nil {
 		t.Fatalf("construct manager: %v", err)
 	}
+	release, err := manager.Admit(context.Background(), "limited")
+	if err != nil {
+		t.Fatalf("fill limited pool: %v", err)
+	}
+	defer release()
 
 	ctx, cancel := context.WithCancel(context.Background())
 	cancel()

internal/defaults/defaults.go

@@ -14,9 +14,6 @@ const (
 	ContentTypeApplicationJSON = "application/json"
 	OpenAIChatCompletionsPath  = "/chat/completions"
 
-	ExecutionDefaultTemperature    = 0.0
-	ExecutionDefaultMaxTokens      = 0
-	ExecutionDefaultTopP           = 1.0
 	ExecutionDefaultTimeoutSeconds = 600
 )
 
@@ -26,9 +23,6 @@ var (
 
 func ExecutionTargetDefault() domain.ExecutionTarget {
 	return domain.ExecutionTarget{
-		Temperature:    ExecutionDefaultTemperature,
-		MaxTokens:      ExecutionDefaultMaxTokens,
-		TopP:           ExecutionDefaultTopP,
 		TimeoutSeconds: ExecutionDefaultTimeoutSeconds,
 	}
 }

internal/domain/domain.go

@@ -145,6 +145,16 @@ type PromptDefinition struct {
 	Validation     OutputContract          `yaml:"validation"`
 }
 
+// PromptInspection is the resolved result of exact prompt inspection.
+type PromptInspection struct {
+	PromptID         string
+	PromptVersion    string
+	PromptHash       string
+	DefaultProfileID string
+	Inputs           []PromptInput
+	OutputContract   OutputContract
+}
+
 // PromptInput describes one named input expected by a prompt definition.
 type PromptInput struct {
 	Name        string `yaml:"name"`
@@ -236,6 +246,13 @@ type ExecutionTarget struct {
 	ExtraParams     map[string]any `yaml:"extra_params" json:"extra_params"`
 }
 
+// ProfileInspection is the resolved result of exact profile inspection.
+type ProfileInspection struct {
+	ProfileID            string
+	EffectiveModelParams ExecutionTarget
+	APIKeyRequired       bool
+}
+
 // OutputContract defines the requirements for the output artifact.
 type OutputContract struct {
 	Format         OutputFormat   `yaml:"format"`

internal/jsonvalue/jsonvalue.go

@@ -1,5 +1,5 @@
 // Package jsonvalue validates and defensively copies JSON-compatible value
-// trees used by public configuration and request boundaries.
+// trees used by configuration, request, and prepared-state boundaries.
 package jsonvalue
 
 import (
@@ -18,13 +18,19 @@ type visit struct {
 	ptr uintptr
 }
 
+// Copy validates and deeply copies a JSON-compatible value while preserving
+// compatible concrete map, slice, array, scalar, and number types.
+func Copy(src any) (any, error) {
+	return copyValue(reflect.ValueOf(src), "value", make(map[visit]struct{}), true)
+}
+
 // CopyMap validates and deeply copies an extra-parameter map while preserving
 // compatible concrete map, slice, array, scalar, and number types.
 func CopyMap(src map[string]any) (map[string]any, error) {
 	if src == nil {
 		return nil, nil
 	}
-	copied, err := copyValue(reflect.ValueOf(src), "extra_params", make(map[visit]struct{}))
+	copied, err := copyValue(reflect.ValueOf(src), "extra_params", make(map[visit]struct{}), false)
 	if err != nil {
 		return nil, err
 	}
@@ -35,7 +41,12 @@ func CopyMap(src map[string]any) (map[string]any, error) {
 	return out, nil
 }
 
-func copyValue(value reflect.Value, path string, seen map[visit]struct{}) (any, error) {
+func copyValue(
+	value reflect.Value,
+	path string,
+	seen map[visit]struct{},
+	allowEmptyMapKeys bool,
+) (any, error) {
 	if !value.IsValid() {
 		return nil, nil
 	}
@@ -43,7 +54,7 @@ func copyValue(value reflect.Value, path string, seen map[visit]struct{}) (any,
 		if value.IsNil() {
 			return nil, nil
 		}
-		return copyValue(value.Elem(), path, seen)
+		return copyValue(value.Elem(), path, seen, allowEmptyMapKeys)
 	}
 	if !value.CanInterface() {
 		return nil, fmt.Errorf("%s: value cannot be copied", path)
@@ -88,22 +99,27 @@ func copyValue(value reflect.Value, path string, seen map[visit]struct{}) (any,
 		}
 		seen[current] = struct{}{}
 		defer delete(seen, current)
-		return copyValue(value.Elem(), path, seen)
+		return copyValue(value.Elem(), path, seen, allowEmptyMapKeys)
 	case reflect.Map:
-		return copyMapValue(value, path, seen)
+		return copyMapValue(value, path, seen, allowEmptyMapKeys)
 	case reflect.Slice:
 		if value.IsNil() {
 			return nil, nil
 		}
-		return copySequenceValue(value, path, seen)
+		return copySequenceValue(value, path, seen, allowEmptyMapKeys)
 	case reflect.Array:
-		return copySequenceValue(value, path, seen)
+		return copySequenceValue(value, path, seen, allowEmptyMapKeys)
 	default:
 		return nil, fmt.Errorf("%s: unsupported JSON value type %s", path, value.Type())
 	}
 }
 
-func copyMapValue(value reflect.Value, path string, seen map[visit]struct{}) (any, error) {
+func copyMapValue(
+	value reflect.Value,
+	path string,
+	seen map[visit]struct{},
+	allowEmptyMapKeys bool,
+) (any, error) {
 	if value.IsNil() {
 		return nil, nil
 	}
@@ -133,10 +149,10 @@ func copyMapValue(value reflect.Value, path string, seen map[visit]struct{}) (an
 	elementType := value.Type().Elem()
 	for _, key := range keys {
 		name := key.String()
-		if name == "" {
+		if name == "" && !allowEmptyMapKeys {
 			return nil, fmt.Errorf("%s: map key must not be empty", path)
 		}
-		copied, err := copyValue(value.MapIndex(key), path+"."+name, seen)
+		copied, err := copyValue(value.MapIndex(key), path+"."+name, seen, allowEmptyMapKeys)
 		if err != nil {
 			return nil, err
 		}
@@ -171,7 +187,12 @@ func copyMapValue(value reflect.Value, path string, seen map[visit]struct{}) (an
 	return out, nil
 }
 
-func copySequenceValue(value reflect.Value, path string, seen map[visit]struct{}) (any, error) {
+func copySequenceValue(
+	value reflect.Value,
+	path string,
+	seen map[visit]struct{},
+	allowEmptyMapKeys bool,
+) (any, error) {
 	var current visit
 	if value.Kind() == reflect.Slice {
 		current = visit{typ: value.Type(), ptr: value.Pointer()}
@@ -186,7 +207,12 @@ func copySequenceValue(value reflect.Value, path string, seen map[visit]struct{}
 	preserveType := true
 	elementType := value.Type().Elem()
 	for i := 0; i < value.Len(); i++ {
-		copied, err := copyValue(value.Index(i), fmt.Sprintf("%s[%d]", path, i), seen)
+		copied, err := copyValue(
+			value.Index(i),
+			fmt.Sprintf("%s[%d]", path, i),
+			seen,
+			allowEmptyMapKeys,
+		)
 		if err != nil {
 			return nil, err
 		}

internal/jsonvalue/jsonvalue_test.go

@@ -44,6 +44,21 @@ func TestCopyMapPreservesTypesAndIsolatesMutations(t *testing.T) {
 	}
 }
 
+func TestCopyAllowsEmptyObjectKeysAndIsolatesMutations(t *testing.T) {
+	nested := map[string]any{"": []any{"original"}}
+
+	copiedValue, err := jsonvalue.Copy(nested)
+	if err != nil {
+		t.Fatalf("copy value: %v", err)
+	}
+	nested[""].([]any)[0] = "changed"
+
+	copied := copiedValue.(map[string]any)
+	if got := copied[""].([]any)[0]; got != "original" {
+		t.Fatalf("copied value was not isolated: %v", got)
+	}
+}
+
 func TestCopyMapRejectsInvalidValues(t *testing.T) {
 	cyclicMap := map[string]any{}
 	cyclicMap["self"] = cyclicMap

internal/profile/builtin/repository.go

@@ -2,7 +2,6 @@ package builtin
 
 import (
 	"embed"
-	"strings"
 
 	"gitea.maximumdirect.net/eric/promptkit/internal/profile"
 )
@@ -15,17 +14,3 @@ var assets embed.FS
 func NewRepository() profile.Repository {
 	return profile.NewFSRepository(assets, assetRoot)
 }
-
-func NewRepositoryWithPrimary(primary profile.Repository) profile.Repository {
-	if primary == nil {
-		return NewRepository()
-	}
-	return profile.NewOverlayRepository(primary, NewRepository())
-}
-
-func NewRepositoryWithDirectory(dir string) profile.Repository {
-	if strings.TrimSpace(dir) == "" {
-		return NewRepository()
-	}
-	return NewRepositoryWithPrimary(profile.NewFilesystemRepository(dir))
-}

internal/profile/builtin/repository_test.go

@@ -2,14 +2,11 @@ package builtin
 
 import (
 	"context"
-	"errors"
 	"io/fs"
 	"strings"
 	"testing"
 
 	"gitea.maximumdirect.net/eric/promptkit/internal/backend"
-	"gitea.maximumdirect.net/eric/promptkit/internal/domain"
-	"gitea.maximumdirect.net/eric/promptkit/internal/profile"
 	"gopkg.in/yaml.v3"
 )
 
@@ -91,53 +88,3 @@ func loadBuiltInProfileIDs(t *testing.T) map[string]string {
 	}
 	return ids
 }
-
-func TestRepositoryWithPrimaryUsesPrimaryBeforeBuiltIns(t *testing.T) {
-	repo := NewRepositoryWithPrimary(staticProfileRepo{
-		profiles: map[string]string{"mistral-small-3": "custom-model"},
-	})
-
-	p, err := repo.GetProfile(context.Background(), "mistral-small-3")
-	if err != nil {
-		t.Fatalf("expected profile to load, got %v", err)
-	}
-	if p.Model != "custom-model" {
-		t.Fatalf("expected primary profile to override built-in, got %+v", p)
-	}
-}
-
-func TestRepositoryWithPrimaryFallsBackToBuiltIns(t *testing.T) {
-	repo := NewRepositoryWithPrimary(staticProfileRepo{})
-
-	p, err := repo.GetProfile(context.Background(), "mistral-small-3")
-	if err != nil {
-		t.Fatalf("expected built-in profile to load, got %v", err)
-	}
-	if p.ID != "mistral-small-3" {
-		t.Fatalf("unexpected profile: %+v", p)
-	}
-}
-
-func TestRepositoryWithPrimaryDoesNotFallBackAfterPrimaryError(t *testing.T) {
-	repo := NewRepositoryWithPrimary(staticProfileRepo{err: profile.ErrInvalidProfile})
-
-	_, err := repo.GetProfile(context.Background(), "mistral-small-3")
-	if !errors.Is(err, profile.ErrInvalidProfile) {
-		t.Fatalf("expected primary error, got %v", err)
-	}
-}
-
-type staticProfileRepo struct {
-	profiles map[string]string
-	err      error
-}
-
-func (r staticProfileRepo) GetProfile(_ context.Context, id string) (*domain.ExecutionProfile, error) {
-	if r.err != nil {
-		return nil, r.err
-	}
-	if model, ok := r.profiles[id]; ok {
-		return &domain.ExecutionProfile{ID: id, Endpoint: "http://primary/v1", Model: model}, nil
-	}
-	return nil, profile.ErrProfileNotFound
-}

internal/usecase/capacity_error.go

@@ -0,0 +1,24 @@
+package usecase
+
+import (
+	"fmt"
+	"strings"
+
+	"gitea.maximumdirect.net/eric/promptkit/internal/capacity"
+)
+
+// CapacityError identifies bounded admission rejected for one selected backend.
+type CapacityError struct {
+	BackendID string
+}
+
+func (e *CapacityError) Error() string {
+	if e == nil || strings.TrimSpace(e.BackendID) == "" {
+		return capacity.ErrCapacityExceeded.Error()
+	}
+	return fmt.Sprintf("backend %q admission: %v", e.BackendID, capacity.ErrCapacityExceeded)
+}
+
+func (e *CapacityError) Unwrap() error {
+	return capacity.ErrCapacityExceeded
+}

internal/usecase/prepared_execution.go

@@ -0,0 +1,298 @@
+package usecase
+
+import (
+	"context"
+	"fmt"
+	"sync"
+	"time"
+
+	"gitea.maximumdirect.net/eric/promptkit/internal/domain"
+	"gitea.maximumdirect.net/eric/promptkit/internal/jsonvalue"
+	"gitea.maximumdirect.net/eric/promptkit/internal/validate"
+)
+
+type preparedExecutionState uint8
+
+const (
+	preparedExecutionReady preparedExecutionState = iota
+	preparedExecutionClaimed
+	preparedExecutionDiscarded
+)
+
+// PreparedExecution owns one frozen, single-use runner execution.
+type PreparedExecution struct {
+	owner   *Runner
+	mu      sync.Mutex
+	state   preparedExecutionState
+	details *domain.PreparedRun
+	payload *preparedExecutionPayload
+}
+
+type preparedExecutionPayload struct {
+	prepared   *domain.PreparedRun
+	validation validate.PreparedValidation
+	directKey  string
+}
+
+// PrepareExecution completes preparation without generation or admission and
+// returns a runner-bound, single-use execution.
+func (r *Runner) PrepareExecution(ctx context.Context, req domain.RunRequest) (*PreparedExecution, error) {
+	state, err := r.resolvePreparation(ctx, req, time.Now().UTC())
+	if err != nil {
+		return nil, err
+	}
+
+	validationPlan, err := r.prepareValidation(ctx, state.effectiveContract)
+	if err != nil {
+		return nil, err
+	}
+
+	structuredOutput, err := r.structuredOutputFromValidationPlan(
+		state.definition,
+		state.effectiveContract,
+		validationPlan,
+	)
+	if err != nil {
+		return nil, err
+	}
+	prepared, err := r.completePreparationWithStructuredOutput(ctx, req, state, structuredOutput)
+	if err != nil {
+		return nil, err
+	}
+
+	executionSnapshot, err := clonePreparedRun(prepared)
+	if err != nil {
+		return nil, fmt.Errorf("%w: failed to copy prepared execution: %v", ErrInvalidRequest, err)
+	}
+
+	details, err := clonePreparedRun(executionSnapshot)
+	if err != nil {
+		return nil, fmt.Errorf("%w: failed to copy prepared execution details: %v", ErrInvalidRequest, err)
+	}
+
+	return &PreparedExecution{
+		owner:   r,
+		state:   preparedExecutionReady,
+		details: details,
+		payload: &preparedExecutionPayload{
+			prepared:   executionSnapshot,
+			validation: validationPlan,
+			directKey:  state.effectiveModel.APIKey,
+		},
+	}, nil
+}
+
+func (r *Runner) prepareValidation(
+	ctx context.Context,
+	contract domain.OutputContract,
+) (validate.PreparedValidation, error) {
+	if r.validator == nil {
+		return noOpPreparedValidation{contract: contract}, nil
+	}
+	preparer, ok := r.validator.(validate.ValidationPreparer)
+	if !ok {
+		return nil, fmt.Errorf("%w: validator does not support prepared validation", ErrValidation)
+	}
+	plan, err := preparer.PrepareValidation(ctx, contract)
+	if err != nil {
+		return nil, fmt.Errorf("%w: %w", ErrValidation, err)
+	}
+	if plan == nil {
+		return nil, fmt.Errorf("%w: validator returned nil prepared validation", ErrValidation)
+	}
+	return plan, nil
+}
+
+func (r *Runner) structuredOutputFromValidationPlan(
+	def *domain.PromptDefinition,
+	contract domain.OutputContract,
+	plan validate.PreparedValidation,
+) (*domain.StructuredOutputSpec, error) {
+	if contract.ValidationMode != domain.ValidationJSONSchema {
+		return nil, nil
+	}
+
+	schemaDocument := plan.SchemaDocument()
+	if schemaDocument == nil {
+		if r.validator == nil {
+			return nil, nil
+		}
+		return nil, fmt.Errorf("%w: prepared json_schema validation has no schema document", ErrValidation)
+	}
+	return structuredOutputSpec(def, schemaDocument), nil
+}
+
+// Details returns a fresh credential-redacted copy of the prepared run.
+func (p *PreparedExecution) Details() *domain.PreparedRun {
+	if p == nil {
+		return nil
+	}
+	p.mu.Lock()
+	detailsSnapshot := p.details
+	p.mu.Unlock()
+
+	details, err := clonePreparedRun(detailsSnapshot)
+	if err != nil {
+		return nil
+	}
+	return details
+}
+
+// Discard invalidates an unclaimed execution and drops its private payload.
+func (p *PreparedExecution) Discard() {
+	if p == nil {
+		return
+	}
+
+	p.mu.Lock()
+	if p.state != preparedExecutionReady {
+		p.mu.Unlock()
+		return
+	}
+	p.state = preparedExecutionDiscarded
+	payload := p.payload
+	p.payload = nil
+	p.mu.Unlock()
+
+	payload.clear()
+}
+
+// RunPrepared claims and executes one prepared execution owned by this runner.
+func (r *Runner) RunPrepared(ctx context.Context, prepared *PreparedExecution) (*domain.RunResult, error) {
+	payload, err := prepared.claim(r)
+	if err != nil {
+		return nil, err
+	}
+	defer payload.clear()
+
+	runID, err := newRunID()
+	if err != nil {
+		return nil, fmt.Errorf("failed to create run id: %w", err)
+	}
+	start := time.Now().UTC()
+
+	target := payload.prepared.EffectiveModelParams
+	if err := validateAPIKey(target.APIKeyEnv, payload.directKey, target.APIKeyRequired); err != nil {
+		return nil, fmt.Errorf("%w: %w", ErrInvalidRequest, err)
+	}
+
+	release, err := r.admitRun(ctx, target.BackendID)
+	if err != nil {
+		return nil, err
+	}
+	defer release()
+
+	return r.executePreparedRun(ctx, payload.prepared, payload.directKey, runID, start, func(
+		ctx context.Context,
+		artifact *domain.Artifact,
+		attemptsUsed int,
+	) (domain.ValidationResult, error) {
+		result, validationErr := payload.validation.Validate(ctx, artifact)
+		if validationErr != nil {
+			return domain.ValidationResult{}, validationErr
+		}
+		result.RepairAttempts = attemptsUsed
+		return result, nil
+	})
+}
+
+func (p *PreparedExecution) claim(owner *Runner) (*preparedExecutionPayload, error) {
+	if p == nil || owner == nil || p.owner != owner {
+		return nil, fmt.Errorf("%w: prepared execution does not belong to this runner", ErrInvalidRequest)
+	}
+
+	p.mu.Lock()
+	defer p.mu.Unlock()
+	if p.state != preparedExecutionReady || p.payload == nil {
+		return nil, fmt.Errorf("%w: prepared execution is not ready", ErrInvalidRequest)
+	}
+	p.state = preparedExecutionClaimed
+	payload := p.payload
+	p.payload = nil
+	return payload, nil
+}
+
+func (p *preparedExecutionPayload) clear() {
+	if p == nil {
+		return
+	}
+	if p.prepared != nil {
+		p.prepared.EffectiveModelParams.APIKey = ""
+	}
+	p.prepared = nil
+	p.validation = nil
+	p.directKey = ""
+}
+
+type noOpPreparedValidation struct {
+	contract domain.OutputContract
+}
+
+func (p noOpPreparedValidation) Validate(
+	ctx context.Context,
+	_ *domain.Artifact,
+) (domain.ValidationResult, error) {
+	if err := ctx.Err(); err != nil {
+		return domain.ValidationResult{}, err
+	}
+	return domain.ValidationResult{
+		Status:         domain.ValidationSkipped,
+		Mode:           p.contract.ValidationMode,
+		SchemaPath:     p.contract.SchemaPath,
+		RepairAttempts: p.contract.RepairAttempts,
+		IsValid:        true,
+	}, nil
+}
+
+func (noOpPreparedValidation) SchemaDocument() any {
+	return nil
+}
+
+func clonePreparedRun(source *domain.PreparedRun) (*domain.PreparedRun, error) {
+	if source == nil {
+		return nil, nil
+	}
+
+	copied := *source
+	extraParams, err := jsonvalue.CopyMap(source.EffectiveModelParams.ExtraParams)
+	if err != nil {
+		return nil, err
+	}
+	copied.EffectiveModelParams.ExtraParams = extraParams
+
+	if source.InputHashes != nil {
+		copied.InputHashes = make(map[string]string, len(source.InputHashes))
+		for name, hash := range source.InputHashes {
+			copied.InputHashes[name] = hash
+		}
+	}
+	if source.Messages != nil {
+		copied.Messages = make([]domain.RenderedMessage, len(source.Messages))
+		for i, message := range source.Messages {
+			copied.Messages[i] = message
+			if message.CacheControl != nil {
+				cacheControl := *message.CacheControl
+				copied.Messages[i].CacheControl = &cacheControl
+			}
+		}
+	}
+
+	if source.StructuredOutput != nil {
+		structuredOutput := *source.StructuredOutput
+		copied.StructuredOutput = &structuredOutput
+		if source.StructuredOutput.JSONSchema != nil {
+			jsonSchema := *source.StructuredOutput.JSONSchema
+			copied.StructuredOutput.JSONSchema = &jsonSchema
+			schema, copyErr := cloneJSONValue(source.StructuredOutput.JSONSchema.Schema)
+			if copyErr != nil {
+				return nil, copyErr
+			}
+			copied.StructuredOutput.JSONSchema.Schema = schema
+		}
+	}
+	return &copied, nil
+}
+
+func cloneJSONValue(source any) (any, error) {
+	return jsonvalue.Copy(source)
+}

internal/usecase/prepared_execution_test.go

@@ -0,0 +1,510 @@
+package usecase
+
+import (
+	"context"
+	"errors"
+	"os"
+	"reflect"
+	"testing"
+
+	"gitea.maximumdirect.net/eric/promptkit/internal/domain"
+	"gitea.maximumdirect.net/eric/promptkit/internal/validate"
+)
+
+type recordingPreparedValidation struct {
+	contract       domain.OutputContract
+	schemaDocument any
+	results        []domain.ValidationResult
+	errs           []error
+	artifacts      []string
+}
+
+func (p *recordingPreparedValidation) Validate(
+	_ context.Context,
+	artifact *domain.Artifact,
+) (domain.ValidationResult, error) {
+	p.artifacts = append(p.artifacts, string(artifact.Body))
+	index := len(p.artifacts) - 1
+	if index < len(p.errs) && p.errs[index] != nil {
+		return domain.ValidationResult{}, p.errs[index]
+	}
+	if len(p.results) == 0 {
+		return domain.ValidationResult{
+			Status:  domain.ValidationPassed,
+			Mode:    p.contract.ValidationMode,
+			IsValid: true,
+		}, nil
+	}
+	if index >= len(p.results) {
+		index = len(p.results) - 1
+	}
+	return p.results[index], nil
+}
+
+func (p *recordingPreparedValidation) SchemaDocument() any {
+	return p.schemaDocument
+}
+
+type recordingValidationPreparer struct {
+	plan                *recordingPreparedValidation
+	prepareErr          error
+	prepareCalls        int
+	directValidateCalls int
+}
+
+func (v *recordingValidationPreparer) Validate(
+	context.Context,
+	*domain.Artifact,
+	domain.OutputContract,
+) (domain.ValidationResult, error) {
+	v.directValidateCalls++
+	return domain.ValidationResult{}, errors.New("live validation must not be used")
+}
+
+func (v *recordingValidationPreparer) PrepareValidation(
+	_ context.Context,
+	contract domain.OutputContract,
+) (validate.PreparedValidation, error) {
+	v.prepareCalls++
+	if v.prepareErr != nil {
+		return nil, v.prepareErr
+	}
+	v.plan.contract = contract
+	return v.plan, nil
+}
+
+func TestRunnerPrepareExecutionCompletesWithoutAdmissionOrGeneration(t *testing.T) {
+	schemaDocument := map[string]any{
+		"type": "object",
+		"properties": map[string]any{
+			"value": map[string]any{"type": "string"},
+			"":      map[string]any{"type": "boolean"},
+		},
+	}
+	def := promptDef(domain.FormatJSON, domain.ValidationJSONSchema, 0)
+	def.Validation.SchemaPath = "schema.json"
+	reader := defaultArtifactReader()
+	renderer := &fakeRenderer{rendered: &domain.RenderedPrompt{
+		SessionID: "prepared-session",
+		Messages: []domain.RenderedMessage{{
+			Role:    "user",
+			Content: "original message",
+		}},
+	}}
+	llmClient := &fakeLLM{forbid: true}
+	validator := &recordingValidationPreparer{
+		plan: &recordingPreparedValidation{schemaDocument: schemaDocument},
+	}
+	admitter := &fakeRunAdmitter{}
+	profile := defaultExecutionProfile()
+	profile.ExtraParams = map[string]any{
+		"metadata": map[string]any{"source": "original"},
+	}
+	runner := NewRunner(
+		&fakePromptRepo{def: def},
+		&fakeExecutionProfileRepo{profiles: map[string]*domain.ExecutionProfile{"exec": profile}},
+		nil,
+		reader,
+		renderer,
+		llmClient,
+		validator,
+		admitter,
+	)
+
+	prepared, err := runner.PrepareExecution(context.Background(), domain.RunRequest{
+		PromptID:  "p",
+		ProfileID: "exec",
+		APIKey:    "direct-test-key",
+		Inputs:    singleInputRef(),
+	})
+	if err != nil {
+		t.Fatalf("prepare execution: %v", err)
+	}
+	defer prepared.Discard()
+
+	if validator.prepareCalls != 1 || validator.directValidateCalls != 0 {
+		t.Fatalf(
+			"validation calls=(prepare=%d direct=%d), want (1, 0)",
+			validator.prepareCalls,
+			validator.directValidateCalls,
+		)
+	}
+	if reader.calls != 1 || renderer.calls != 1 {
+		t.Fatalf("completion calls=(artifact=%d render=%d), want (1, 1)", reader.calls, renderer.calls)
+	}
+	if len(admitter.backendIDs) != 0 || llmClient.calls != 0 {
+		t.Fatalf("prepare invoked execution collaborators: admission=%v generation=%d", admitter.backendIDs, llmClient.calls)
+	}
+
+	first := prepared.Details()
+	if first == nil {
+		t.Fatal("prepared details are nil")
+	}
+	if first.EffectiveModelParams.APIKey != "" {
+		t.Fatal("prepared details retained the direct API key")
+	}
+	if first.StructuredOutput == nil ||
+		first.StructuredOutput.JSONSchema == nil ||
+		!reflect.DeepEqual(first.StructuredOutput.JSONSchema.Schema, schemaDocument) {
+		t.Fatalf("prepared details have unexpected structured output: %#v", first.StructuredOutput)
+	}
+
+	first.Messages[0].Content = "caller mutation"
+	first.InputHashes["input"] = "caller mutation"
+	first.EffectiveModelParams.ExtraParams["metadata"].(map[string]any)["source"] = "caller mutation"
+	first.StructuredOutput.JSONSchema.Schema.(map[string]any)["type"] = "string"
+	renderer.rendered.Messages[0].Content = "source mutation"
+
+	second := prepared.Details()
+	if second.Messages[0].Content != "original message" ||
+		second.InputHashes["input"] == "caller mutation" ||
+		second.EffectiveModelParams.ExtraParams["metadata"].(map[string]any)["source"] != "original" ||
+		second.StructuredOutput.JSONSchema.Schema.(map[string]any)["type"] != "object" {
+		t.Fatalf("details did not preserve an independent snapshot: %#v", second)
+	}
+}
+
+func TestRunnerRunPreparedRechecksEnvironmentCredentialBeforeAdmission(t *testing.T) {
+	const environmentName = "PROMPTKIT_PREPARED_EXECUTION_TEST_KEY"
+	t.Setenv(environmentName, "available-during-preparation")
+
+	profile := defaultExecutionProfile()
+	profile.APIKeyEnv = environmentName
+	validator := &recordingValidationPreparer{plan: &recordingPreparedValidation{}}
+	admitter := &fakeRunAdmitter{}
+	llmClient := &fakeLLM{resp: &domain.GenerateResponse{Content: "unexpected"}}
+	runner := NewRunner(
+		&fakePromptRepo{def: promptDef(domain.FormatText, domain.ValidationNone, 0)},
+		&fakeExecutionProfileRepo{profiles: map[string]*domain.ExecutionProfile{"exec": profile}},
+		nil,
+		defaultArtifactReader(),
+		defaultRenderer(),
+		llmClient,
+		validator,
+		admitter,
+	)
+	prepared, err := runner.PrepareExecution(context.Background(), domain.RunRequest{
+		PromptID:  "p",
+		ProfileID: "exec",
+		Inputs:    singleInputRef(),
+	})
+	if err != nil {
+		t.Fatalf("prepare execution: %v", err)
+	}
+	if err := os.Unsetenv(environmentName); err != nil {
+		t.Fatalf("unset credential environment: %v", err)
+	}
+
+	result, err := runner.RunPrepared(context.Background(), prepared)
+	if result != nil {
+		t.Fatalf("credential failure returned partial result: %+v", result)
+	}
+	if !errors.Is(err, ErrInvalidRequest) || !errors.Is(err, ErrAPIKeyEnvMissing) {
+		t.Fatalf("credential error identities are missing: %v", err)
+	}
+	if len(admitter.backendIDs) != 0 || llmClient.calls != 0 {
+		t.Fatalf("credential failure reached admission or generation: admission=%v generation=%d", admitter.backendIDs, llmClient.calls)
+	}
+	if _, err := runner.RunPrepared(context.Background(), prepared); !errors.Is(err, ErrInvalidRequest) {
+		t.Fatalf("credential failure did not consume execution: %v", err)
+	}
+}
+
+func TestRunnerRunPreparedKeepsDirectCredentialOutOfMetadata(t *testing.T) {
+	const directKey = "direct-prepared-test-key"
+
+	profile := defaultExecutionProfile()
+	profile.APIKeyRequired = true
+	client := &fakeLLM{resp: &domain.GenerateResponse{Content: "ok"}}
+	runner := NewRunner(
+		&fakePromptRepo{def: promptDef(domain.FormatText, domain.ValidationNone, 0)},
+		&fakeExecutionProfileRepo{profiles: map[string]*domain.ExecutionProfile{"exec": profile}},
+		nil,
+		defaultArtifactReader(),
+		defaultRenderer(),
+		client,
+		&recordingValidationPreparer{plan: &recordingPreparedValidation{}},
+		nil,
+	)
+
+	prepared, err := runner.PrepareExecution(context.Background(), domain.RunRequest{
+		PromptID:  "p",
+		ProfileID: "exec",
+		APIKey:    directKey,
+		Inputs:    singleInputRef(),
+	})
+	if err != nil {
+		t.Fatalf("prepare execution: %v", err)
+	}
+	if details := prepared.Details(); details.EffectiveModelParams.APIKey != "" {
+		t.Fatal("prepared details retained direct credential")
+	}
+
+	result, err := runner.RunPrepared(context.Background(), prepared)
+	if err != nil {
+		t.Fatalf("run prepared: %v", err)
+	}
+	if client.lastReq.Target.APIKey != directKey {
+		t.Fatal("generation did not receive direct credential")
+	}
+	if result.EffectiveModelParams.APIKey != "" {
+		t.Fatal("run result retained direct credential")
+	}
+}
+
+func TestRunnerRunPreparedUsesFrozenValidationForInitialAndRepairOutputs(t *testing.T) {
+	validator := &recordingValidationPreparer{
+		plan: &recordingPreparedValidation{
+			results: []domain.ValidationResult{
+				{
+					Status:  domain.ValidationFailed,
+					Mode:    domain.ValidationJSON,
+					Errors:  []string{"invalid"},
+					IsValid: false,
+				},
+				{
+					Status:  domain.ValidationPassed,
+					Mode:    domain.ValidationJSON,
+					IsValid: true,
+				},
+			},
+		},
+	}
+	repairer := &fakeRepairer{
+		responses: []*domain.GenerateResponse{{Content: `{"repaired":true}`}},
+	}
+	admitter := &fakeRunAdmitter{}
+	reader := defaultArtifactReader()
+	renderer := defaultRenderer()
+	runner := NewRunnerWithRepairer(
+		&fakePromptRepo{def: promptDef(domain.FormatJSON, domain.ValidationJSON, 1)},
+		&fakeExecutionProfileRepo{profiles: map[string]*domain.ExecutionProfile{"exec": defaultExecutionProfile()}},
+		nil,
+		reader,
+		renderer,
+		&fakeLLM{resp: &domain.GenerateResponse{Content: `{"broken":true}`}},
+		validator,
+		repairer,
+		admitter,
+	)
+	prepared, err := runner.PrepareExecution(context.Background(), domain.RunRequest{
+		PromptID:  "p",
+		ProfileID: "exec",
+		Inputs:    singleInputRef(),
+	})
+	if err != nil {
+		t.Fatalf("prepare execution: %v", err)
+	}
+
+	result, err := runner.RunPrepared(context.Background(), prepared)
+	if err != nil {
+		t.Fatalf("run prepared: %v", err)
+	}
+	if !reflect.DeepEqual(validator.plan.artifacts, []string{`{"broken":true}`, `{"repaired":true}`}) {
+		t.Fatalf("prepared validation artifacts=%#v", validator.plan.artifacts)
+	}
+	if validator.directValidateCalls != 0 || repairer.calls != 1 {
+		t.Fatalf("validation/repair calls=(direct=%d repair=%d), want (0, 1)", validator.directValidateCalls, repairer.calls)
+	}
+	if result.Validation.Status != domain.ValidationPassed || result.Validation.RepairAttempts != 1 {
+		t.Fatalf("unexpected repaired validation result: %+v", result.Validation)
+	}
+	if admitter.releaseCalls != 1 {
+		t.Fatalf("admission releases=%d, want 1", admitter.releaseCalls)
+	}
+	if reader.calls != 1 || renderer.calls != 1 {
+		t.Fatalf("execution reopened preparation sources: artifact=%d render=%d", reader.calls, renderer.calls)
+	}
+}
+
+func TestRunnerRunPreparedReleasesAdmissionAcrossExecutionErrors(t *testing.T) {
+	generationFailure := errors.New("generation failed")
+	validationFailure := errors.New("validation failed")
+	repairFailure := errors.New("repair failed")
+
+	tests := []struct {
+		name          string
+		generationErr error
+		validation    *recordingPreparedValidation
+		repairer      *fakeRepairer
+		wantError     error
+	}{
+		{
+			name:          "generation failure",
+			generationErr: generationFailure,
+			validation:    &recordingPreparedValidation{},
+			wantError:     ErrLLMGenerate,
+		},
+		{
+			name: "validation failure",
+			validation: &recordingPreparedValidation{
+				errs: []error{validationFailure},
+			},
+			wantError: ErrValidation,
+		},
+		{
+			name: "repair failure",
+			validation: &recordingPreparedValidation{
+				results: []domain.ValidationResult{{
+					Status:  domain.ValidationFailed,
+					Mode:    domain.ValidationJSON,
+					Errors:  []string{"invalid"},
+					IsValid: false,
+				}},
+			},
+			repairer:  &fakeRepairer{err: repairFailure},
+			wantError: ErrValidation,
+		},
+	}
+
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			def := promptDef(domain.FormatJSON, domain.ValidationJSON, 1)
+			if test.repairer == nil {
+				def.Validation.RepairAttempts = 0
+			}
+			validator := &recordingValidationPreparer{plan: test.validation}
+			admitter := &fakeRunAdmitter{}
+			runner := NewRunnerWithRepairer(
+				&fakePromptRepo{def: def},
+				&fakeExecutionProfileRepo{profiles: map[string]*domain.ExecutionProfile{"exec": defaultExecutionProfile()}},
+				nil,
+				defaultArtifactReader(),
+				defaultRenderer(),
+				&fakeLLM{
+					resp: &domain.GenerateResponse{Content: `{"value":true}`},
+					err:  test.generationErr,
+				},
+				validator,
+				test.repairer,
+				admitter,
+			)
+			prepared, err := runner.PrepareExecution(context.Background(), domain.RunRequest{
+				PromptID:  "p",
+				ProfileID: "exec",
+				Inputs:    singleInputRef(),
+			})
+			if err != nil {
+				t.Fatalf("prepare execution: %v", err)
+			}
+
+			result, err := runner.RunPrepared(context.Background(), prepared)
+			if result != nil || !errors.Is(err, test.wantError) {
+				t.Fatalf("run prepared=(%+v, %v), want %v", result, err, test.wantError)
+			}
+			if len(admitter.backendIDs) != 1 || admitter.releaseCalls != 1 {
+				t.Fatalf(
+					"admission calls=%#v releases=%d, want one each",
+					admitter.backendIDs,
+					admitter.releaseCalls,
+				)
+			}
+		})
+	}
+}
+
+func TestRunnerPreparedExecutionOwnershipUseAndDiscard(t *testing.T) {
+	newRunner := func() *Runner {
+		return NewRunner(
+			&fakePromptRepo{def: promptDef(domain.FormatText, domain.ValidationNone, 0)},
+			&fakeExecutionProfileRepo{profiles: map[string]*domain.ExecutionProfile{"exec": defaultExecutionProfile()}},
+			nil,
+			defaultArtifactReader(),
+			defaultRenderer(),
+			&fakeLLM{resp: &domain.GenerateResponse{Content: "ok"}},
+			&recordingValidationPreparer{plan: &recordingPreparedValidation{}},
+			nil,
+		)
+	}
+	request := domain.RunRequest{
+		PromptID:  "p",
+		ProfileID: "exec",
+		Inputs:    singleInputRef(),
+	}
+
+	owner := newRunner()
+	prepared, err := owner.PrepareExecution(context.Background(), request)
+	if err != nil {
+		t.Fatalf("prepare execution: %v", err)
+	}
+	if _, err := newRunner().RunPrepared(context.Background(), prepared); !errors.Is(err, ErrInvalidRequest) {
+		t.Fatalf("foreign runner error=%v, want ErrInvalidRequest", err)
+	}
+	if _, err := owner.RunPrepared(context.Background(), prepared); err != nil {
+		t.Fatalf("owner run prepared: %v", err)
+	}
+	if _, err := owner.RunPrepared(context.Background(), prepared); !errors.Is(err, ErrInvalidRequest) {
+		t.Fatalf("second owner run error=%v, want ErrInvalidRequest", err)
+	}
+	if prepared.Details() == nil {
+		t.Fatal("details unavailable after execution")
+	}
+
+	discarded, err := owner.PrepareExecution(context.Background(), request)
+	if err != nil {
+		t.Fatalf("prepare discarded execution: %v", err)
+	}
+	discarded.Discard()
+	discarded.Discard()
+	if _, err := owner.RunPrepared(context.Background(), discarded); !errors.Is(err, ErrInvalidRequest) {
+		t.Fatalf("discarded execution error=%v, want ErrInvalidRequest", err)
+	}
+	if discarded.Details() == nil {
+		t.Fatal("details unavailable after discard")
+	}
+}
+
+func TestRunnerPreparedExecutionWithoutValidatorSkipsValidation(t *testing.T) {
+	runner := NewRunner(
+		&fakePromptRepo{def: promptDef(domain.FormatJSON, domain.ValidationJSON, 0)},
+		&fakeExecutionProfileRepo{profiles: map[string]*domain.ExecutionProfile{"exec": defaultExecutionProfile()}},
+		nil,
+		defaultArtifactReader(),
+		defaultRenderer(),
+		&fakeLLM{resp: &domain.GenerateResponse{Content: `{}`}},
+		nil,
+		nil,
+	)
+	prepared, err := runner.PrepareExecution(context.Background(), domain.RunRequest{
+		PromptID:  "p",
+		ProfileID: "exec",
+		Inputs:    singleInputRef(),
+	})
+	if err != nil {
+		t.Fatalf("prepare execution: %v", err)
+	}
+
+	result, err := runner.RunPrepared(context.Background(), prepared)
+	if err != nil {
+		t.Fatalf("run prepared: %v", err)
+	}
+	if result.Validation.Status != domain.ValidationSkipped || !result.Validation.IsValid {
+		t.Fatalf("unexpected no-validator result: %+v", result.Validation)
+	}
+}
+
+func TestRunnerPrepareExecutionRequiresValidationPreparer(t *testing.T) {
+	reader := defaultArtifactReader()
+	runner := NewRunner(
+		&fakePromptRepo{def: promptDef(domain.FormatText, domain.ValidationBasic, 0)},
+		&fakeExecutionProfileRepo{profiles: map[string]*domain.ExecutionProfile{"exec": defaultExecutionProfile()}},
+		nil,
+		reader,
+		defaultRenderer(),
+		&fakeLLM{forbid: true},
+		&fakeValidator{},
+		nil,
+	)
+
+	prepared, err := runner.PrepareExecution(context.Background(), domain.RunRequest{
+		PromptID:  "p",
+		ProfileID: "exec",
+		Inputs:    singleInputRef(),
+	})
+	if prepared != nil || !errors.Is(err, ErrValidation) {
+		t.Fatalf("prepare execution=(%+v, %v), want ErrValidation", prepared, err)
+	}
+	if reader.calls != 0 {
+		t.Fatalf("unsupported validator allowed completion, artifact calls=%d", reader.calls)
+	}
+}

internal/usecase/profile_inspection.go

@@ -0,0 +1,103 @@
+package usecase
+
+import (
+	"context"
+	"errors"
+	"fmt"
+	"strings"
+
+	"gitea.maximumdirect.net/eric/promptkit/internal/domain"
+)
+
+type resolvedProfileSelection struct {
+	id      string
+	profile *domain.ExecutionProfile
+	backend *domain.Backend
+}
+
+func (r *Runner) resolveProfileSelection(
+	ctx context.Context,
+	profileID string,
+) (*resolvedProfileSelection, error) {
+	normalizedID := strings.TrimSpace(profileID)
+	if normalizedID == "" {
+		return nil, fmt.Errorf("%w: profile id is required", ErrInvalidRequest)
+	}
+	if r == nil || r.profiles == nil {
+		return nil, fmt.Errorf("%w: profile repository is not configured", ErrProfileLoad)
+	}
+
+	selectedProfile, err := r.profiles.GetProfile(ctx, normalizedID)
+	if err != nil {
+		return nil, fmt.Errorf("%w: %w", ErrProfileLoad, err)
+	}
+	if selectedProfile == nil {
+		return nil, fmt.Errorf("%w: profile repository returned nil profile", ErrProfileLoad)
+	}
+
+	profileValue := *selectedProfile
+	profileValue.BackendID = strings.TrimSpace(profileValue.BackendID)
+
+	var selectedBackend *domain.Backend
+	if profileValue.BackendID != "" {
+		if r.backends == nil {
+			return nil, fmt.Errorf("%w: backend %q cannot be resolved", ErrProfileLoad, profileValue.BackendID)
+		}
+		backendValue, err := r.backends.GetBackend(profileValue.BackendID)
+		if err != nil {
+			return nil, fmt.Errorf("%w: backend %q: %w", ErrProfileLoad, profileValue.BackendID, err)
+		}
+		selectedBackend = &backendValue
+	}
+
+	return &resolvedProfileSelection{
+		id:      normalizedID,
+		profile: &profileValue,
+		backend: selectedBackend,
+	}, nil
+}
+
+func validateResolvedExecutionTarget(target domain.ExecutionTarget) error {
+	if strings.TrimSpace(target.Endpoint) == "" {
+		return errors.New("execution endpoint is required")
+	}
+	if strings.TrimSpace(target.Model) == "" {
+		return errors.New("execution model is required")
+	}
+	return nil
+}
+
+// InspectProfile resolves one explicit profile without prompt or execution work.
+func (r *Runner) InspectProfile(
+	ctx context.Context,
+	profileID string,
+) (*domain.ProfileInspection, error) {
+	normalizedID := strings.TrimSpace(profileID)
+	if normalizedID == "" {
+		return nil, fmt.Errorf("%w: profile id is required", ErrInvalidRequest)
+	}
+	select {
+	case <-ctx.Done():
+		return nil, fmt.Errorf("%w: %w", ErrProfileLoad, ctx.Err())
+	default:
+	}
+
+	selection, err := r.resolveProfileSelection(ctx, normalizedID)
+	if err != nil {
+		return nil, err
+	}
+	target, _, err := resolveExecutionTarget(selection.backend, selection.profile, nil)
+	if err != nil {
+		return nil, fmt.Errorf("%w: %w", ErrProfileLoad, err)
+	}
+	if err := validateResolvedExecutionTarget(target); err != nil {
+		return nil, fmt.Errorf("%w: %w", ErrProfileLoad, err)
+	}
+	target.APIKey = ""
+
+	return &domain.ProfileInspection{
+		ProfileID:            selection.id,
+		EffectiveModelParams: target,
+		APIKeyRequired:       target.APIKeyRequired,
+	}, nil
+}

internal/usecase/profile_inspection_test.go

@@ -0,0 +1,213 @@
+package usecase
+
+import (
+	"context"
+	"errors"
+	"reflect"
+	"testing"
+
+	"gitea.maximumdirect.net/eric/promptkit/internal/defaults"
+	"gitea.maximumdirect.net/eric/promptkit/internal/domain"
+	"gitea.maximumdirect.net/eric/promptkit/internal/profile"
+)
+
+type inspectionProfileRepository struct {
+	profile *domain.ExecutionProfile
+	err     error
+	calls   int
+	id      string
+}
+
+func (r *inspectionProfileRepository) GetProfile(
+	_ context.Context,
+	id string,
+) (*domain.ExecutionProfile, error) {
+	r.calls++
+	r.id = id
+	if r.err != nil {
+		return nil, r.err
+	}
+	return r.profile, nil
+}
+
+type inspectionBackendResolver struct {
+	backend domain.Backend
+	err     error
+	calls   int
+	id      string
+}
+
+func (r *inspectionBackendResolver) GetBackend(id string) (domain.Backend, error) {
+	r.calls++
+	r.id = id
+	if r.err != nil {
+		return domain.Backend{}, r.err
+	}
+	return r.backend, nil
+}
+
+func TestRunnerInspectProfileResolvesProfileAndBackendOnce(t *testing.T) {
+	profiles := &inspectionProfileRepository{profile: &domain.ExecutionProfile{
+		ID:              "profile",
+		BackendID:       "  backend  ",
+		Model:           "profile-model",
+		Temperature:     0.4,
+		MaxTokens:       32,
+		TimeoutSeconds:  45,
+		ServiceTier:     "priority",
+		ReasoningEffort: "high",
+		ExtraParams: map[string]any{
+			"profile": "value",
+		},
+	}}
+	backends := &inspectionBackendResolver{backend: domain.Backend{
+		ID:        "backend",
+		Endpoint:  "https://backend.example/v1",
+		APIKeyEnv: "BACKEND_KEY",
+		ExtraParams: map[string]any{
+			"backend": "value",
+		},
+	}}
+	runner := &Runner{profiles: profiles, backends: backends}
+
+	inspection, err := runner.InspectProfile(context.Background(), " profile ")
+	if err != nil {
+		t.Fatalf("inspect profile: %v", err)
+	}
+	if profiles.calls != 1 || profiles.id != "profile" {
+		t.Fatalf("profile lookup=(calls=%d id=%q), want one exact lookup", profiles.calls, profiles.id)
+	}
+	if backends.calls != 1 || backends.id != "backend" {
+		t.Fatalf("backend lookup=(calls=%d id=%q), want one exact lookup", backends.calls, backends.id)
+	}
+	if profiles.profile.BackendID != "  backend  " {
+		t.Fatalf("inspection mutated repository profile backend: %q", profiles.profile.BackendID)
+	}
+
+	wantTarget := domain.ExecutionTarget{
+		BackendID:       "backend",
+		Endpoint:        "https://backend.example/v1",
+		Model:           "profile-model",
+		Temperature:     0.4,
+		MaxTokens:       32,
+		TopP:            defaults.ExecutionTargetDefault().TopP,
+		TimeoutSeconds:  45,
+		ServiceTier:     "priority",
+		ReasoningEffort: "high",
+		APIKeyEnv:       "BACKEND_KEY",
+		ExtraParams: map[string]any{
+			"profile": "value",
+		},
+	}
+	if inspection.ProfileID != "profile" || inspection.APIKeyRequired ||
+		!reflect.DeepEqual(inspection.EffectiveModelParams, wantTarget) {
+		t.Fatalf("inspection=%#v, want profile=%q target=%#v", inspection, "profile", wantTarget)
+	}
+}
+
+func TestRunnerInspectProfileDoesNotNeedExecutionCollaboratorsOrCredentials(t *testing.T) {
+	t.Setenv("PROMPTKIT_INSPECTION_TEST_KEY", "")
+	profiles := &inspectionProfileRepository{profile: &domain.ExecutionProfile{
+		ID:        "endpoint-only",
+		Endpoint:  "https://profile.example/v1",
+		Model:     "profile-model",
+		APIKeyEnv: "PROMPTKIT_INSPECTION_TEST_KEY",
+	}}
+	runner := &Runner{profiles: profiles}
+
+	inspection, err := runner.InspectProfile(context.Background(), "endpoint-only")
+	if err != nil {
+		t.Fatalf("inspect endpoint-only profile: %v", err)
+	}
+	if inspection.EffectiveModelParams.BackendID != "" ||
+		inspection.EffectiveModelParams.APIKeyEnv != "PROMPTKIT_INSPECTION_TEST_KEY" ||
+		inspection.APIKeyRequired {
+		t.Fatalf("unexpected endpoint-only inspection: %#v", inspection)
+	}
+}
+
+func TestRunnerInspectProfileDirectCredentialRequirementClearsBackendEnvironment(t *testing.T) {
+	profiles := &inspectionProfileRepository{profile: &domain.ExecutionProfile{
+		ID:             "direct-key",
+		BackendID:      "backend",
+		Model:          "profile-model",
+		APIKeyRequired: true,
+	}}
+	backends := &inspectionBackendResolver{backend: domain.Backend{
+		ID:        "backend",
+		Endpoint:  "https://backend.example/v1",
+		APIKeyEnv: "BACKEND_KEY",
+	}}
+
+	inspection, err := (&Runner{profiles: profiles, backends: backends}).InspectProfile(
+		context.Background(),
+		"direct-key",
+	)
+	if err != nil {
+		t.Fatalf("inspect direct-key profile: %v", err)
+	}
+	if !inspection.APIKeyRequired || inspection.EffectiveModelParams.APIKeyEnv != "" {
+		t.Fatalf("credential requirement was not resolved exclusively: %#v", inspection)
+	}
+}
+
+func TestRunnerInspectProfileClassifiesFailuresWithoutRepositoryWorkAfterCancellation(t *testing.T) {
+	t.Run("blank ID", func(t *testing.T) {
+		profiles := &inspectionProfileRepository{}
+		_, err := (&Runner{profiles: profiles}).InspectProfile(context.Background(), " \t ")
+		if !errors.Is(err, ErrInvalidRequest) || profiles.calls != 0 {
+			t.Fatalf("blank inspection=(%v, calls=%d), want invalid request without lookup", err, profiles.calls)
+		}
+	})
+
+	t.Run("canceled context", func(t *testing.T) {
+		ctx, cancel := context.WithCancel(context.Background())
+		cancel()
+		profiles := &inspectionProfileRepository{}
+		_, err := (&Runner{profiles: profiles}).InspectProfile(ctx, "profile")
+		if !errors.Is(err, ErrProfileLoad) || !errors.Is(err, context.Canceled) || profiles.calls != 0 {
+			t.Fatalf("canceled inspection=(%v, calls=%d), want profile load and context identities without lookup", err, profiles.calls)
+		}
+	})
+
+	t.Run("missing profile", func(t *testing.T) {
+		profiles := &inspectionProfileRepository{err: profile.ErrProfileNotFound}
+		_, err := (&Runner{profiles: profiles}).InspectProfile(context.Background(), "missing")
+		if !errors.Is(err, ErrProfileLoad) || !errors.Is(err, profile.ErrProfileNotFound) {
+			t.Fatalf("missing profile error=%v, want profile load and not-found identities", err)
+		}
+	})
+
+	t.Run("unknown backend", func(t *testing.T) {
+		backendErr := errors.New("unknown backend")
+		profiles := &inspectionProfileRepository{profile: &domain.ExecutionProfile{
+			ID: "profile", BackendID: "backend", Model: "profile-model",
+		}}
+		backends := &inspectionBackendResolver{err: backendErr}
+		_, err := (&Runner{profiles: profiles, backends: backends}).InspectProfile(context.Background(), "profile")
+		if !errors.Is(err, ErrProfileLoad) || !errors.Is(err, backendErr) {
+			t.Fatalf("unknown backend error=%v, want profile load and backend identities", err)
+		}
+	})
+
+	t.Run("defensive invalid dependencies", func(t *testing.T) {
+		cases := []struct {
+			name   string
+			runner *Runner
+		}{
+			{name: "nil repository", runner: &Runner{}},
+			{name: "nil profile", runner: &Runner{profiles: &inspectionProfileRepository{}}},
+			{name: "invalid target", runner: &Runner{profiles: &inspectionProfileRepository{
+				profile: &domain.ExecutionProfile{ID: "profile", Endpoint: "https://profile.example/v1"},
+			}}},
+		}
+		for _, tc := range cases {
+			t.Run(tc.name, func(t *testing.T) {
+				_, err := tc.runner.InspectProfile(context.Background(), "profile")
+				if !errors.Is(err, ErrProfileLoad) {
+					t.Fatalf("inspection error=%v, want profile load", err)
+				}
+			})
+		}
+	})
+}

internal/usecase/prompt_inspection.go

@@ -0,0 +1,73 @@
+package usecase
+
+import (
+	"context"
+	"fmt"
+	"strings"
+
+	"gitea.maximumdirect.net/eric/promptkit/internal/domain"
+)
+
+type resolvedPromptDefinition struct {
+	definition *domain.PromptDefinition
+	hash       string
+}
+
+func (r *Runner) resolvePromptDefinition(
+	ctx context.Context,
+	promptID string,
+	promptVersion string,
+) (*resolvedPromptDefinition, error) {
+	if strings.TrimSpace(promptID) == "" {
+		return nil, fmt.Errorf("%w: prompt id is required", ErrInvalidRequest)
+	}
+	if r == nil || r.promptDefs == nil {
+		return nil, fmt.Errorf("%w: prompt repository is not configured", ErrPromptLoad)
+	}
+
+	definition, err := r.promptDefs.GetPromptDefinition(ctx, promptID, promptVersion)
+	if err != nil {
+		return nil, fmt.Errorf("%w: %w", ErrPromptLoad, err)
+	}
+	if definition == nil {
+		return nil, fmt.Errorf("%w: prompt repository returned nil definition", ErrPromptLoad)
+	}
+
+	hash, err := hashPromptDefinition(definition)
+	if err != nil {
+		return nil, fmt.Errorf("%w: failed to hash prompt definition: %v", ErrPromptLoad, err)
+	}
+	return &resolvedPromptDefinition{definition: definition, hash: hash}, nil
+}
+
+// InspectPrompt resolves one explicit prompt without execution work.
+func (r *Runner) InspectPrompt(
+	ctx context.Context,
+	promptID string,
+	promptVersion string,
+) (*domain.PromptInspection, error) {
+	if strings.TrimSpace(promptID) == "" {
+		return nil, fmt.Errorf("%w: prompt id is required", ErrInvalidRequest)
+	}
+	select {
+	case <-ctx.Done():
+		return nil, fmt.Errorf("%w: %w", ErrPromptLoad, ctx.Err())
+	default:
+	}
+
+	selection, err := r.resolvePromptDefinition(ctx, promptID, promptVersion)
+	if err != nil {
+		return nil, err
+	}
+	inputs := make([]domain.PromptInput, len(selection.definition.Inputs))
+	copy(inputs, selection.definition.Inputs)
+
+	return &domain.PromptInspection{
+		PromptID:         selection.definition.ID,
+		PromptVersion:    selection.definition.Version,
+		PromptHash:       selection.hash,
+		DefaultProfileID: selection.definition.DefaultProfile,
+		Inputs:           inputs,
+		OutputContract:   selection.definition.Validation,
+	}, nil
+}

internal/usecase/prompt_inspection_test.go

@@ -0,0 +1,157 @@
+package usecase
+
+import (
+	"context"
+	"errors"
+	"reflect"
+	"testing"
+
+	"gitea.maximumdirect.net/eric/promptkit/internal/domain"
+	"gitea.maximumdirect.net/eric/promptkit/internal/promptdef"
+)
+
+type inspectionPromptRepository struct {
+	definition *domain.PromptDefinition
+	err        error
+	calls      int
+	id         string
+	version    string
+}
+
+func (r *inspectionPromptRepository) GetPromptDefinition(
+	_ context.Context,
+	id string,
+	version string,
+) (*domain.PromptDefinition, error) {
+	r.calls++
+	r.id = id
+	r.version = version
+	if r.err != nil {
+		return nil, r.err
+	}
+	return r.definition, nil
+}
+
+func TestRunnerInspectPromptResolvesOneDefinitionWithoutExecutionCollaborators(t *testing.T) {
+	definition := &domain.PromptDefinition{
+		ID:             "normalized.prompt",
+		Version:        "1.2.3",
+		DefaultProfile: "not-resolved",
+		Inputs: []domain.PromptInput{
+			{Name: "document", Required: true, ContentType: "text/plain", Description: "Source document."},
+			{Name: "audience", ContentType: "text/plain", Description: "Intended reader."},
+		},
+		Validation: domain.OutputContract{
+			Format:         domain.FormatJSON,
+			ValidationMode: domain.ValidationJSONSchema,
+			SchemaPath:     "schemas/result.json",
+			RepairAttempts: 2,
+		},
+	}
+	repository := &inspectionPromptRepository{definition: definition}
+	runner := &Runner{promptDefs: repository}
+
+	inspection, err := runner.InspectPrompt(context.Background(), " prompt-id ", " version ")
+	if err != nil {
+		t.Fatalf("inspect prompt: %v", err)
+	}
+	wantHash, err := hashPromptDefinition(definition)
+	if err != nil {
+		t.Fatalf("hash prompt definition: %v", err)
+	}
+	if repository.calls != 1 || repository.id != " prompt-id " || repository.version != " version " {
+		t.Fatalf("prompt lookup=(calls=%d id=%q version=%q), want one unchanged lookup", repository.calls, repository.id, repository.version)
+	}
+	if inspection.PromptID != definition.ID ||
+		inspection.PromptVersion != definition.Version ||
+		inspection.PromptHash != wantHash ||
+		inspection.DefaultProfileID != definition.DefaultProfile ||
+		!reflect.DeepEqual(inspection.Inputs, definition.Inputs) ||
+		inspection.OutputContract != definition.Validation {
+		t.Fatalf("inspection=%#v, want definition metadata", inspection)
+	}
+
+	inspection.Inputs[0].Name = "changed"
+	second, err := runner.InspectPrompt(context.Background(), " prompt-id ", " version ")
+	if err != nil {
+		t.Fatalf("inspect prompt again: %v", err)
+	}
+	if definition.Inputs[0].Name != "document" || second.Inputs[0].Name != "document" {
+		t.Fatalf("inspection input mutation escaped caller result: definition=%#v next=%#v", definition.Inputs, second.Inputs)
+	}
+}
+
+func TestRunnerInspectPromptClassifiesFailuresWithoutRepositoryWorkAfterCancellation(t *testing.T) {
+	t.Run("blank ID", func(t *testing.T) {
+		repository := &inspectionPromptRepository{}
+		_, err := (&Runner{promptDefs: repository}).InspectPrompt(context.Background(), " \t ", "version")
+		if !errors.Is(err, ErrInvalidRequest) || repository.calls != 0 {
+			t.Fatalf("blank inspection=(%v, calls=%d), want invalid request without lookup", err, repository.calls)
+		}
+	})
+
+	t.Run("canceled context", func(t *testing.T) {
+		ctx, cancel := context.WithCancel(context.Background())
+		cancel()
+		repository := &inspectionPromptRepository{}
+		_, err := (&Runner{promptDefs: repository}).InspectPrompt(ctx, "prompt", "version")
+		if !errors.Is(err, ErrPromptLoad) || !errors.Is(err, context.Canceled) || repository.calls != 0 {
+			t.Fatalf("canceled inspection=(%v, calls=%d), want prompt load and context identities without lookup", err, repository.calls)
+		}
+	})
+
+	t.Run("missing prompt", func(t *testing.T) {
+		repository := &inspectionPromptRepository{err: promptdef.ErrPromptDefinitionNotFound}
+		_, err := (&Runner{promptDefs: repository}).InspectPrompt(context.Background(), "missing", "version")
+		if !errors.Is(err, ErrPromptLoad) || !errors.Is(err, promptdef.ErrPromptDefinitionNotFound) {
+			t.Fatalf("missing prompt error=%v, want prompt load and not-found identities", err)
+		}
+	})
+
+	t.Run("defensive prompt dependencies", func(t *testing.T) {
+		var nilRunner *Runner
+		cases := []struct {
+			name   string
+			runner *Runner
+		}{
+			{name: "nil runner", runner: nilRunner},
+			{name: "nil repository", runner: &Runner{}},
+			{name: "nil definition", runner: &Runner{promptDefs: &inspectionPromptRepository{}}},
+		}
+		for _, tc := range cases {
+			t.Run(tc.name, func(t *testing.T) {
+				_, err := tc.runner.InspectPrompt(context.Background(), "prompt", "version")
+				if !errors.Is(err, ErrPromptLoad) {
+					t.Fatalf("inspection error=%v, want prompt load", err)
+				}
+			})
+		}
+	})
+}
+
+func TestRunnerPrepareUsesThePromptInspectionSelectionAndHash(t *testing.T) {
+	definition := promptDef(domain.FormatMarkdown, domain.ValidationBasic, 0)
+	repository := &fakePromptRepo{def: definition}
+	runner := NewRunner(
+		repository,
+		&fakeExecutionProfileRepo{profiles: map[string]*domain.ExecutionProfile{"exec": defaultExecutionProfile()}},
+		nil,
+		&fakeArtifactReader{},
+		&fakeRenderer{rendered: &domain.RenderedPrompt{Messages: []domain.RenderedMessage{{Role: "user", Content: "hello"}}}},
+		nil,
+		nil,
+		nil,
+	)
+
+	inspection, err := runner.InspectPrompt(context.Background(), definition.ID, definition.Version)
+	if err != nil {
+		t.Fatalf("inspect prompt: %v", err)
+	}
+	prepared, err := runner.Prepare(context.Background(), domain.RunRequest{PromptID: definition.ID, PromptVersion: definition.Version, ProfileID: "exec"})
+	if err != nil {
+		t.Fatalf("prepare prompt: %v", err)
+	}
+	if inspection.PromptHash != prepared.PromptHash {
+		t.Fatalf("inspection hash=%q, preparation hash=%q", inspection.PromptHash, prepared.PromptHash)
+	}
+}

internal/usecase/runner.go

@@ -131,29 +131,46 @@ func (r *Runner) Run(ctx context.Context, req domain.RunRequest) (*domain.RunRes
 		return nil, err
 	}
 
-	if r.admitter != nil {
-		release, admitErr := r.admitter.Admit(ctx, state.effectiveModel.BackendID)
-		if admitErr != nil {
-			if errors.Is(admitErr, capacity.ErrCapacityExceeded) {
-				return nil, fmt.Errorf(
-					"backend %q admission: %w",
-					state.effectiveModel.BackendID,
-					admitErr,
-				)
-			}
-			return nil, admitErr
-		}
-		defer release()
+	release, err := r.admitRun(ctx, state.effectiveModel.BackendID)
+	if err != nil {
+		return nil, err
 	}
+	defer release()
 
 	prepared, err := r.completePreparation(ctx, req, state)
 	if err != nil {
 		return nil, err
 	}
+	directAPIKey := state.effectiveModel.APIKey
 
+	return r.executePreparedRun(ctx, prepared, directAPIKey, runID, start, func(
+		ctx context.Context,
+		artifact *domain.Artifact,
+		attemptsUsed int,
+	) (domain.ValidationResult, error) {
+		return r.validateOutput(ctx, artifact, prepared.OutputContract, attemptsUsed)
+	})
+}
+
+type preparedValidationFunc func(
+	context.Context,
+	*domain.Artifact,
+	int,
+) (domain.ValidationResult, error)
+
+func (r *Runner) executePreparedRun(
+	ctx context.Context,
+	prepared *domain.PreparedRun,
+	directAPIKey string,
+	runID string,
+	start time.Time,
+	validateArtifact preparedValidationFunc,
+) (*domain.RunResult, error) {
+	executionTarget := prepared.EffectiveModelParams
+	executionTarget.APIKey = directAPIKey
 	genResp, err := r.llm.Generate(ctx, domain.GenerateRequest{
 		Prompt:           domain.RenderedPrompt{SessionID: prepared.SessionID, Messages: prepared.Messages},
-		Target:           prepared.EffectiveModelParams,
+		Target:           executionTarget,
 		TargetPresence:   prepared.TargetPresence,
 		StructuredOutput: prepared.StructuredOutput,
 	})
@@ -165,7 +182,7 @@ func (r *Runner) Run(ctx context.Context, req domain.RunRequest) (*domain.RunRes
 	}
 
 	outputArtifact := buildOutputArtifact(genResp.Content, prepared.OutputContract.Format)
-	validationResult, err := r.validateOutput(ctx, &outputArtifact, prepared.OutputContract, 0)
+	validationResult, err := validateArtifact(ctx, &outputArtifact, 0)
 	if err != nil {
 		return nil, fmt.Errorf("%w: %w", ErrValidation, err)
 	}
@@ -179,7 +196,7 @@ func (r *Runner) Run(ctx context.Context, req domain.RunRequest) (*domain.RunRes
 				PreviousOutput:   genResp.Content,
 				ValidationErrors: validationResult.Errors,
 				SessionID:        prepared.SessionID,
-				Target:           prepared.EffectiveModelParams,
+				Target:           executionTarget,
 				StructuredOutput: prepared.StructuredOutput,
 				Attempt:          attemptsUsed,
 				MaxAttempts:      prepared.OutputContract.RepairAttempts,
@@ -195,7 +212,7 @@ func (r *Runner) Run(ctx context.Context, req domain.RunRequest) (*domain.RunRes
 			genResp = repairResp
 			outputArtifact = buildOutputArtifact(genResp.Content, prepared.OutputContract.Format)
 
-			validationResult, err = r.validateOutput(ctx, &outputArtifact, prepared.OutputContract, attemptsUsed)
+			validationResult, err = validateArtifact(ctx, &outputArtifact, attemptsUsed)
 			if err != nil {
 				return nil, fmt.Errorf("%w: %w", ErrValidation, err)
 			}
@@ -203,6 +220,7 @@ func (r *Runner) Run(ctx context.Context, req domain.RunRequest) (*domain.RunRes
 	}
 
 	end := time.Now().UTC()
+	executionTarget.APIKey = ""
 
 	return &domain.RunResult{
 		RunID:                runID,
@@ -218,7 +236,7 @@ func (r *Runner) Run(ctx context.Context, req domain.RunRequest) (*domain.RunRes
 		SelectedBackendID:    prepared.SelectedBackendID,
 		ModelName:            prepared.EffectiveModelParams.Model,
 		Endpoint:             prepared.EffectiveModelParams.Endpoint,
-		EffectiveModelParams: prepared.EffectiveModelParams,
+		EffectiveModelParams: executionTarget,
 		InputHashes:          prepared.InputHashes,
 		Usage:                genResp.Usage,
 		StartTime:            start,
@@ -248,14 +266,12 @@ func (r *Runner) resolvePreparation(
 		return nil, fmt.Errorf("%w: session_id: %v", ErrInvalidRequest, err)
 	}
 
-	def, err := r.promptDefs.GetPromptDefinition(ctx, req.PromptID, req.PromptVersion)
+	promptSelection, err := r.resolvePromptDefinition(ctx, req.PromptID, req.PromptVersion)
 	if err != nil {
-		return nil, fmt.Errorf("%w: %w", ErrPromptLoad, err)
-	}
-	promptDefinitionHash, err := hashPromptDefinition(def)
-	if err != nil {
-		return nil, fmt.Errorf("%w: failed to hash prompt definition: %v", ErrPromptLoad, err)
+		return nil, err
 	}
+	def := promptSelection.definition
+	promptDefinitionHash := promptSelection.hash
 
 	selectedProfileID := strings.TrimSpace(req.ProfileID)
 	if selectedProfileID == "" {
@@ -265,34 +281,18 @@ func (r *Runner) resolvePreparation(
 		return nil, fmt.Errorf("%w: %w: profile id is required either in request or prompt default_profile", ErrInvalidRequest, ErrProfileRequired)
 	}
 
-	execProfile, err := r.profiles.GetProfile(ctx, selectedProfileID)
+	selection, err := r.resolveProfileSelection(ctx, selectedProfileID)
 	if err != nil {
-		return nil, fmt.Errorf("%w: %w", ErrProfileLoad, err)
+		return nil, err
 	}
 
-	var selectedBackend *domain.Backend
-	if backendID := strings.TrimSpace(execProfile.BackendID); backendID != "" {
-		execProfile.BackendID = backendID
-		if r.backends == nil {
-			return nil, fmt.Errorf("%w: backend %q cannot be resolved", ErrProfileLoad, backendID)
-		}
-		resolvedBackend, resolveErr := r.backends.GetBackend(backendID)
-		if resolveErr != nil {
-			return nil, fmt.Errorf("%w: backend %q: %w", ErrProfileLoad, backendID, resolveErr)
-		}
-		selectedBackend = &resolvedBackend
-	}
-
-	effectiveModel, targetPresence, err := resolveExecutionTarget(selectedBackend, execProfile, req.Execution)
+	effectiveModel, targetPresence, err := resolveExecutionTarget(selection.backend, selection.profile, req.Execution)
 	if err != nil {
 		return nil, fmt.Errorf("%w: %w", ErrInvalidRequest, err)
 	}
 	effectiveModel.APIKey = req.APIKey
-	if strings.TrimSpace(effectiveModel.Endpoint) == "" {
-		return nil, fmt.Errorf("%w: execution endpoint is required", ErrInvalidRequest)
-	}
-	if strings.TrimSpace(effectiveModel.Model) == "" {
-		return nil, fmt.Errorf("%w: execution model is required", ErrInvalidRequest)
+	if err := validateResolvedExecutionTarget(effectiveModel); err != nil {
+		return nil, fmt.Errorf("%w: %w", ErrInvalidRequest, err)
 	}
 	if err := validateAPIKey(effectiveModel.APIKeyEnv, effectiveModel.APIKey, effectiveModel.APIKeyRequired); err != nil {
 		return nil, fmt.Errorf("%w: %w", ErrInvalidRequest, err)
@@ -303,7 +303,7 @@ func (r *Runner) resolvePreparation(
 		definition:           def,
 		directSessionID:      directSessionID,
 		promptDefinitionHash: promptDefinitionHash,
-		selectedProfileID:    selectedProfileID,
+		selectedProfileID:    selection.id,
 		effectiveModel:       effectiveModel,
 		targetPresence:       targetPresence,
 		effectiveContract:    effectiveContract,
@@ -324,7 +324,15 @@ func (r *Runner) completePreparation(
 	if err != nil {
 		return nil, err
 	}
+	return r.completePreparationWithStructuredOutput(ctx, req, state, structuredOutput)
+}
 
+func (r *Runner) completePreparationWithStructuredOutput(
+	ctx context.Context,
+	req domain.RunRequest,
+	state *preparationState,
+	structuredOutput *domain.StructuredOutputSpec,
+) (*domain.PreparedRun, error) {
 	resolvedInputs := make(map[string]*domain.Artifact, len(req.Inputs))
 	inputHashes := make(map[string]string, len(req.Inputs))
 	for name, ref := range req.Inputs {
@@ -354,13 +362,15 @@ func (r *Runner) completePreparation(
 	}
 
 	end := time.Now().UTC()
+	effectiveModel := state.effectiveModel
+	effectiveModel.APIKey = ""
 	return &domain.PreparedRun{
 		PromptID:             state.definition.ID,
 		PromptVersion:        state.definition.Version,
 		PromptHash:           state.promptDefinitionHash,
 		SelectedProfileID:    state.selectedProfileID,
 		SelectedBackendID:    state.effectiveModel.BackendID,
-		EffectiveModelParams: state.effectiveModel,
+		EffectiveModelParams: effectiveModel,
 		TargetPresence:       state.targetPresence,
 		OutputContract:       state.effectiveContract,
 		StructuredOutput:     structuredOutput,
@@ -374,6 +384,20 @@ func (r *Runner) completePreparation(
 	}, nil
 }
 
+func (r *Runner) admitRun(ctx context.Context, backendID string) (func(), error) {
+	if r.admitter == nil {
+		return func() {}, nil
+	}
+	release, err := r.admitter.Admit(ctx, backendID)
+	if err != nil {
+		if errors.Is(err, capacity.ErrCapacityExceeded) && strings.TrimSpace(backendID) != "" {
+			return nil, &CapacityError{BackendID: backendID}
+		}
+		return nil, err
+	}
+	return release, nil
+}
+
 func (r *Runner) resolveStructuredOutput(ctx context.Context, def *domain.PromptDefinition, contract domain.OutputContract) (*domain.StructuredOutputSpec, error) {
 	if contract.ValidationMode != domain.ValidationJSONSchema {
 		return nil, nil
@@ -389,14 +413,18 @@ func (r *Runner) resolveStructuredOutput(ctx context.Context, def *domain.Prompt
 		return nil, fmt.Errorf("%w: failed to load json schema for structured output: %v", ErrValidation, err)
 	}
 
+	return structuredOutputSpec(def, schemaDoc), nil
+}
+
+func structuredOutputSpec(def *domain.PromptDefinition, schemaDocument any) *domain.StructuredOutputSpec {
 	return &domain.StructuredOutputSpec{
 		Type: domain.StructuredOutputJSONSchema,
 		JSONSchema: &domain.StructuredOutputJSONSpec{
 			Name:   deriveStructuredSchemaName(def.ID, def.Version),
 			Strict: true,
-			Schema: schemaDoc,
+			Schema: schemaDocument,
 		},
-	}, nil
+	}
 }
 
 func deriveStructuredSchemaName(promptID string, promptVersion string) string {

internal/usecase/runner_test.go

@@ -1357,14 +1357,14 @@ func TestRunnerAdmissionUsesResolvedBackendIdentity(t *testing.T) {
 
 func TestRunnerAdmissionFailureSkipsCompletionCollaborators(t *testing.T) {
 	tests := []struct {
-		name               string
-		admissionError     error
-		wantBackendContext bool
+		name             string
+		admissionError   error
+		wantCapacityType bool
 	}{
 		{
-			name:               "capacity exhausted",
-			admissionError:     capacity.ErrCapacityExceeded,
-			wantBackendContext: true,
+			name:             "capacity exhausted",
+			admissionError:   capacity.ErrCapacityExceeded,
+			wantCapacityType: true,
 		},
 		{
 			name:           "context canceled",
@@ -1414,8 +1414,16 @@ func TestRunnerAdmissionFailureSkipsCompletionCollaborators(t *testing.T) {
 			if errors.Is(err, ErrInvalidRequest) || errors.Is(err, ErrLLMGenerate) {
 				t.Fatalf("admission error was recategorized: %v", err)
 			}
-			if tc.wantBackendContext && !strings.Contains(err.Error(), "custom") {
-				t.Fatalf("capacity error lacks backend context: %v", err)
+			var capacityErr *CapacityError
+			if tc.wantCapacityType {
+				if !errors.As(err, &capacityErr) {
+					t.Fatalf("capacity error=%v, want internal typed identity", err)
+				}
+				if capacityErr.BackendID != "custom" {
+					t.Fatalf("capacity backend ID=%q, want custom", capacityErr.BackendID)
+				}
+			} else if errors.As(err, &capacityErr) {
+				t.Fatalf("non-capacity admission error exposed typed capacity identity: %v", err)
 			}
 			if !reflect.DeepEqual(admitter.backendIDs, []string{"custom"}) {
 				t.Fatalf("admitted backend IDs=%#v, want custom", admitter.backendIDs)
@@ -1682,36 +1690,6 @@ func TestRunnerRunSelectedProfileBeatsBuiltInDefault(t *testing.T) {
 	}
 }
 
-func TestRunnerRunBuiltInDefaultsUsedWhenProfileOmitsOptionalFields(t *testing.T) {
-	promptRepo := &fakePromptRepo{def: promptDef(domain.FormatText, domain.ValidationNone, 0)}
-	execRepo := &fakeExecutionProfileRepo{profiles: map[string]*domain.ExecutionProfile{
-		"exec": {ID: "exec", Endpoint: "http://profile/v1", Model: "profile-model"},
-	}}
-	llmClient := &fakeLLM{resp: &domain.GenerateResponse{Content: "ok"}}
-	runner := NewRunner(promptRepo, execRepo, nil, defaultArtifactReader(), defaultRenderer(), llmClient, nil, nil)
-
-	res, err := runner.Run(context.Background(), domain.RunRequest{
-		PromptID:  "p",
-		ProfileID: "exec",
-		Inputs:    singleInputRef(),
-	})
-	if err != nil {
-		t.Fatalf("expected no error, got %v", err)
-	}
-	if res.EffectiveModelParams.Temperature != defaults.ExecutionDefaultTemperature {
-		t.Fatalf("expected default temperature %v, got %v", defaults.ExecutionDefaultTemperature, res.EffectiveModelParams.Temperature)
-	}
-	if res.EffectiveModelParams.TopP != defaults.ExecutionDefaultTopP {
-		t.Fatalf("expected default top_p %v, got %v", defaults.ExecutionDefaultTopP, res.EffectiveModelParams.TopP)
-	}
-	if res.EffectiveModelParams.MaxTokens != defaults.ExecutionDefaultMaxTokens {
-		t.Fatalf("expected default max_tokens %d, got %d", defaults.ExecutionDefaultMaxTokens, res.EffectiveModelParams.MaxTokens)
-	}
-	if res.EffectiveModelParams.TimeoutSeconds != defaults.ExecutionDefaultTimeoutSeconds {
-		t.Fatalf("expected default timeout_seconds %d, got %d", defaults.ExecutionDefaultTimeoutSeconds, res.EffectiveModelParams.TimeoutSeconds)
-	}
-}
-
 func TestRunnerRunAPIKeyEnvResolvesFromEnvironment(t *testing.T) {
 	t.Setenv("PROMPTKIT_TEST_API_KEY", "secret")
 	promptRepo := &fakePromptRepo{def: promptDef(domain.FormatText, domain.ValidationNone, 0)}
@@ -1757,7 +1735,7 @@ func TestRunnerRunDirectAPIKeyBypassesMissingEnvAndReachesLLM(t *testing.T) {
 	llmClient := &fakeLLM{resp: &domain.GenerateResponse{Content: "ok"}}
 	runner := NewRunner(promptRepo, execRepo, nil, defaultArtifactReader(), defaultRenderer(), llmClient, nil, nil)
 
-	_, err := runner.Run(context.Background(), domain.RunRequest{
+	result, err := runner.Run(context.Background(), domain.RunRequest{
 		PromptID:  "p",
 		ProfileID: "exec",
 		APIKey:    directKey,
@@ -1769,6 +1747,9 @@ func TestRunnerRunDirectAPIKeyBypassesMissingEnvAndReachesLLM(t *testing.T) {
 	if llmClient.lastReq.Target.APIKey != directKey {
 		t.Fatalf("expected direct API key to reach LLM request")
 	}
+	if result.EffectiveModelParams.APIKey != "" {
+		t.Fatal("run result retained direct API key")
+	}
 	if llmClient.lastReq.Target.APIKeyEnv != "PROMPTKIT_MISSING_KEY" {
 		t.Fatalf("expected api_key_env name to remain on target, got %q", llmClient.lastReq.Target.APIKeyEnv)
 	}

internal/validate/standard_validator.go

@@ -45,6 +45,101 @@ func (v *FSValidator) Validate(ctx context.Context, artifact *domain.Artifact, c
 	return validateArtifact(ctx, artifact, contract, v.validateJSONSchema)
 }
 
+type preparedValidation struct {
+	contract       domain.OutputContract
+	schemaDocument any
+	schema         *jsonschema.Schema
+}
+
+func (p *preparedValidation) Validate(ctx context.Context, artifact *domain.Artifact) (domain.ValidationResult, error) {
+	return validateArtifact(ctx, artifact, p.contract, p.validateJSONSchema)
+}
+
+func (p *preparedValidation) SchemaDocument() any {
+	return p.schemaDocument
+}
+
+func (p *preparedValidation) validateJSONSchema(instance any, _ string) ([]string, error) {
+	if p.schema == nil {
+		return nil, errors.New("prepared JSON schema is unavailable")
+	}
+	if err := p.schema.Validate(instance); err != nil {
+		return []string{fmt.Sprintf("json schema validation failed: %v", err)}, nil
+	}
+	return nil, nil
+}
+
+func (v *StandardValidator) PrepareValidation(ctx context.Context, contract domain.OutputContract) (PreparedValidation, error) {
+	if err := ctx.Err(); err != nil {
+		return nil, err
+	}
+
+	prepared := &preparedValidation{contract: contract}
+	if contract.ValidationMode != domain.ValidationJSONSchema {
+		return prepared, nil
+	}
+
+	resolvedSchemaPath, err := v.resolveSchemaPath(contract.SchemaPath)
+	if err != nil {
+		return nil, err
+	}
+	schemaDocument, err := loadJSONSchemaFile(resolvedSchemaPath)
+	if err != nil {
+		return nil, fmt.Errorf("failed to compile JSON schema %q: %w", resolvedSchemaPath, err)
+	}
+	schemaRoot, err := v.schemaRoot()
+	if err != nil {
+		return nil, err
+	}
+	compiler := newSchemaCompiler(standardSchemaLoader{root: schemaRoot})
+	if err := compiler.AddResource(resolvedSchemaPath, schemaDocument); err != nil {
+		return nil, fmt.Errorf("failed to register JSON schema %q: %w", resolvedSchemaPath, err)
+	}
+	schema, err := compiler.Compile(resolvedSchemaPath)
+	if err != nil {
+		return nil, fmt.Errorf("failed to compile JSON schema %q: %w", resolvedSchemaPath, err)
+	}
+	if err := ctx.Err(); err != nil {
+		return nil, err
+	}
+
+	prepared.schemaDocument = schemaDocument
+	prepared.schema = schema
+	return prepared, nil
+}
+
+func (v *FSValidator) PrepareValidation(ctx context.Context, contract domain.OutputContract) (PreparedValidation, error) {
+	if err := ctx.Err(); err != nil {
+		return nil, err
+	}
+
+	prepared := &preparedValidation{contract: contract}
+	if contract.ValidationMode != domain.ValidationJSONSchema {
+		return prepared, nil
+	}
+
+	schemaName, schemaDocument, err := v.loadSchemaDocument(contract.SchemaPath)
+	if err != nil {
+		return nil, err
+	}
+	resourceURL := fsSchemaResourceURL(schemaName)
+	compiler := newSchemaCompiler(fsSchemaLoader{fsys: v.fsys, root: filecatalog.CleanFSRoot(v.root)})
+	if err := compiler.AddResource(resourceURL, schemaDocument); err != nil {
+		return nil, fmt.Errorf("failed to register JSON schema %q: %w", schemaName, err)
+	}
+	schema, err := compiler.Compile(resourceURL)
+	if err != nil {
+		return nil, fmt.Errorf("failed to compile JSON schema %q: %w", schemaName, err)
+	}
+	if err := ctx.Err(); err != nil {
+		return nil, err
+	}
+
+	prepared.schemaDocument = schemaDocument
+	prepared.schema = schema
+	return prepared, nil
+}
+
 type schemaValidatorFunc func(instance any, schemaPath string) ([]string, error)
 
 func validateArtifact(ctx context.Context, artifact *domain.Artifact, contract domain.OutputContract, validateSchema schemaValidatorFunc) (domain.ValidationResult, error) {

internal/validate/standard_validator_test.go

@@ -5,6 +5,7 @@ import (
 	"encoding/json"
 	"os"
 	"path/filepath"
+	"reflect"
 	"strconv"
 	"strings"
 	"testing"
@@ -120,6 +121,68 @@ func TestStandardValidatorJSONSchemaSuccess(t *testing.T) {
 	}
 }
 
+func TestStandardValidatorPreparedSchemaSurvivesSourceRemoval(t *testing.T) {
+	tmp := t.TempDir()
+	rootSchema := []byte(`{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "title": "original root",
+  "type": "object",
+  "required": ["value"],
+  "properties": {
+    "value": {"$ref": "value.json"}
+  }
+}`)
+	rootPath := filepath.Join(tmp, "schema.json")
+	referencePath := filepath.Join(tmp, "value.json")
+	if err := os.WriteFile(rootPath, rootSchema, 0o644); err != nil {
+		t.Fatal(err)
+	}
+	if err := os.WriteFile(referencePath, []byte(`{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "type": "integer",
+  "minimum": 2
+}`), 0o644); err != nil {
+		t.Fatal(err)
+	}
+
+	validator := NewStandardValidator(tmp)
+	preparer, ok := validator.(ValidationPreparer)
+	if !ok {
+		t.Fatal("standard validator does not support validation preparation")
+	}
+	prepared, err := preparer.PrepareValidation(context.Background(), domain.OutputContract{
+		ValidationMode: domain.ValidationJSONSchema,
+		SchemaPath:     "schema.json",
+	})
+	if err != nil {
+		t.Fatalf("prepare validation: %v", err)
+	}
+	assertSchemaDocument(t, prepared.SchemaDocument(), rootSchema)
+
+	if err := os.Remove(rootPath); err != nil {
+		t.Fatal(err)
+	}
+	if err := os.Remove(referencePath); err != nil {
+		t.Fatal(err)
+	}
+
+	valid, err := prepared.Validate(context.Background(), &domain.Artifact{Body: []byte(`{"value":3}`)})
+	if err != nil {
+		t.Fatalf("validate prepared artifact: %v", err)
+	}
+	if valid.Status != domain.ValidationPassed || !valid.IsValid {
+		t.Fatalf("expected passed/valid, got status=%q valid=%v errors=%v", valid.Status, valid.IsValid, valid.Errors)
+	}
+
+	invalid, err := prepared.Validate(context.Background(), &domain.Artifact{Body: []byte(`{"value":"changed"}`)})
+	if err != nil {
+		t.Fatalf("validate prepared artifact: %v", err)
+	}
+	if invalid.Status != domain.ValidationFailed || invalid.IsValid {
+		t.Fatalf("expected failed/invalid, got status=%q valid=%v", invalid.Status, invalid.IsValid)
+	}
+}
+
 func TestStandardValidatorJSONSchemaNestedSchemaPathSuccess(t *testing.T) {
 	tmp := t.TempDir()
 	nestedDir := filepath.Join(tmp, "dnd")
@@ -295,6 +358,64 @@ func TestFSValidatorJSONSchemaSuccess(t *testing.T) {
 	}
 }
 
+func TestFSValidatorPreparedSchemaSurvivesSourceMutation(t *testing.T) {
+	rootSchema := []byte(`{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "title": "original root",
+  "type": "object",
+  "required": ["value"],
+  "properties": {
+    "value": {"$ref": "value.json"}
+  }
+}`)
+	fsys := fstest.MapFS{
+		"schema.json": &fstest.MapFile{Data: rootSchema},
+		"value.json": &fstest.MapFile{Data: []byte(`{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "type": "integer",
+  "minimum": 2
+}`)},
+	}
+	validator := NewFSValidator(fsys, ".")
+	preparer, ok := validator.(ValidationPreparer)
+	if !ok {
+		t.Fatal("filesystem validator does not support validation preparation")
+	}
+	prepared, err := preparer.PrepareValidation(context.Background(), domain.OutputContract{
+		ValidationMode: domain.ValidationJSONSchema,
+		SchemaPath:     "schema.json",
+	})
+	if err != nil {
+		t.Fatalf("prepare validation: %v", err)
+	}
+	assertSchemaDocument(t, prepared.SchemaDocument(), rootSchema)
+
+	fsys["schema.json"] = &fstest.MapFile{Data: []byte(`{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "type": "string"
+}`)}
+	fsys["value.json"] = &fstest.MapFile{Data: []byte(`{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "type": "string"
+}`)}
+
+	valid, err := prepared.Validate(context.Background(), &domain.Artifact{Body: []byte(`{"value":3}`)})
+	if err != nil {
+		t.Fatalf("validate prepared artifact: %v", err)
+	}
+	if valid.Status != domain.ValidationPassed || !valid.IsValid {
+		t.Fatalf("expected passed/valid, got status=%q valid=%v errors=%v", valid.Status, valid.IsValid, valid.Errors)
+	}
+
+	invalid, err := prepared.Validate(context.Background(), &domain.Artifact{Body: []byte(`{"value":"changed"}`)})
+	if err != nil {
+		t.Fatalf("validate prepared artifact: %v", err)
+	}
+	if invalid.Status != domain.ValidationFailed || invalid.IsValid {
+		t.Fatalf("expected failed/invalid, got status=%q valid=%v", invalid.Status, invalid.IsValid)
+	}
+}
+
 func TestFSValidatorJSONSchemaRegistrationError(t *testing.T) {
 	v := NewFSValidator(fstest.MapFS{
 		"schemas/%zz.json": &fstest.MapFile{Data: []byte(`{"type":"object"}`)},
@@ -548,3 +669,15 @@ func TestJSONSchemaDialectIsDraft2020(t *testing.T) {
 		})
 	}
 }
+
+func assertSchemaDocument(t *testing.T, got any, expectedJSON []byte) {
+	t.Helper()
+
+	var expected any
+	if err := json.Unmarshal(expectedJSON, &expected); err != nil {
+		t.Fatalf("decode expected schema document: %v", err)
+	}
+	if !reflect.DeepEqual(got, expected) {
+		t.Fatalf("schema document mismatch:\n got: %#v\nwant: %#v", got, expected)
+	}
+}

internal/validate/validator.go

@@ -2,6 +2,7 @@ package validate
 
 import (
 	"context"
+
 	"gitea.maximumdirect.net/eric/promptkit/internal/domain"
 )
 
@@ -10,6 +11,20 @@ type Validator interface {
 	Validate(ctx context.Context, artifact *domain.Artifact, contract domain.OutputContract) (domain.ValidationResult, error)
 }
 
+// PreparedValidation validates artifacts against one frozen output contract.
+type PreparedValidation interface {
+	Validate(ctx context.Context, artifact *domain.Artifact) (domain.ValidationResult, error)
+	// SchemaDocument returns the root JSON Schema document used for provider
+	// structured output, or nil for non-schema modes. Returned internal
+	// immutable state must not be mutated.
+	SchemaDocument() any
+}
+
+// ValidationPreparer freezes validation resources for one output contract.
+type ValidationPreparer interface {
+	PrepareValidation(ctx context.Context, contract domain.OutputContract) (PreparedValidation, error)
+}
+
 // SchemaDocumentLoader loads JSON schema documents using validator path semantics.
 type SchemaDocumentLoader interface {
 	LoadSchemaDocument(ctx context.Context, schemaPath string) (any, error)

prepared_execution.go

@@ -0,0 +1,55 @@
+package promptkit
+
+import "gitea.maximumdirect.net/eric/promptkit/internal/usecase"
+
+const preparedExecutionString = "promptkit.PreparedExecution{opaque}"
+
+// PreparedExecution is an opaque, in-process handle for one completely
+// prepared execution. A handle is bound to the [Engine] that created it and
+// permits one [Engine.RunPrepared] invocation.
+//
+// PreparedExecution contains no supported serializable state and cannot be
+// used as a restartable job. Copying the value preserves the same shared
+// lifecycle; it does not create another execution attempt.
+type PreparedExecution struct {
+	internal *usecase.PreparedExecution
+}
+
+// Details returns a fresh caller-owned, credential-redacted copy of the
+// prepared request details. Mutating the result cannot affect execution or a
+// later Details call. Details remains available after execution or discard.
+//
+// A nil receiver or zero-value PreparedExecution returns a zero [PreparedRun].
+func (p *PreparedExecution) Details() PreparedRun {
+	if p == nil || p.internal == nil {
+		return PreparedRun{}
+	}
+	details := fromDomainPreparedRun(p.internal.Details())
+	if details == nil {
+		return PreparedRun{}
+	}
+	return *details
+}
+
+// Discard invalidates an unclaimed handle and drops Promptkit's references to
+// its execution-only state. Discard is nil-safe and idempotent. It does not
+// cancel an execution that has already claimed the handle; use the
+// [Engine.RunPrepared] context for cancellation.
+func (p *PreparedExecution) Discard() {
+	if p == nil || p.internal == nil {
+		return
+	}
+	p.internal.Discard()
+}
+
+// String returns a constant representation that exposes no retained request,
+// rendered content, or credential data.
+func (p *PreparedExecution) String() string {
+	return preparedExecutionString
+}
+
+// GoString returns a constant Go-syntax representation that exposes no
+// retained request, rendered content, or credential data.
+func (p *PreparedExecution) GoString() string {
+	return preparedExecutionString
+}

prepared_execution_contract_test.go

@@ -0,0 +1,773 @@
+package promptkit_test
+
+import (
+	"context"
+	"encoding/json"
+	"errors"
+	"fmt"
+	"os"
+	"reflect"
+	"strings"
+	"sync"
+	"testing"
+	"testing/fstest"
+	"time"
+
+	"gitea.maximumdirect.net/eric/promptkit"
+)
+
+func TestPreparedExecutionFreezesSourcesAndReturnsIndependentDetails(t *testing.T) {
+	promptSource := preparedPromptSource("original")
+	profileSource := preparedProfileSource("original-model")
+	schemaSource := preparedSchemaSource()
+	reader := &mutablePreparedArtifactReader{
+		body: "original artifact",
+		hash: "original-input-hash",
+	}
+	client := &preparedRecordingClient{
+		response: &promptkit.GenerateResponse{Content: `{"value":3}`},
+	}
+	engine, err := promptkit.NewEngine(
+		promptkit.Config{},
+		promptkit.WithPromptFS(promptSource, "."),
+		promptkit.WithProfileFS(profileSource, "."),
+		promptkit.WithSchemaFS(schemaSource, "."),
+		promptkit.WithArtifactReader(reader),
+		promptkit.WithLLMClient(client),
+	)
+	if err != nil {
+		t.Fatalf("construct engine: %v", err)
+	}
+
+	temperature := 0.25
+	extraParams := map[string]any{
+		"nested": map[string]any{"source": "original"},
+	}
+	request := promptkit.RunRequest{
+		PromptID: "prepared",
+		Inputs: map[string]promptkit.ArtifactRef{
+			"input": promptkit.Inline("original request input"),
+		},
+		Vars: map[string]string{"label": "original variable"},
+		Execution: &promptkit.ExecutionTargetOverride{
+			Temperature: &temperature,
+			ExtraParams: extraParams,
+		},
+	}
+	preparationContext, cancelPreparation := context.WithCancel(context.Background())
+	prepared, err := engine.PrepareExecution(preparationContext, request)
+	if err != nil {
+		t.Fatalf("prepare execution: %v", err)
+	}
+	cancelPreparation()
+
+	request.PromptID = "changed"
+	request.Inputs["input"] = promptkit.Inline("changed request input")
+	request.Vars["label"] = "changed variable"
+	temperature = 1.5
+	extraParams["nested"].(map[string]any)["source"] = "changed"
+	promptSource["prompt.yaml"] = &fstest.MapFile{Data: []byte(`id: changed`)}
+	profileSource["profile.yaml"] = &fstest.MapFile{Data: []byte(`id: changed`)}
+	schemaSource["schema.json"] = &fstest.MapFile{Data: []byte(`{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "title": "changed root",
+  "type": "string"
+}`)}
+	schemaSource["value.json"] = &fstest.MapFile{Data: []byte(`{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "type": "string"
+}`)}
+	reader.set("changed artifact", "changed-input-hash")
+
+	first := prepared.Details()
+	first.Messages[0].Content = "changed details"
+	first.InputHashes["input"] = "changed-details-hash"
+	first.EffectiveModelParams.ExtraParams["nested"].(map[string]any)["source"] = "changed details"
+	first.StructuredOutput.JSONSchema.Schema.(map[string]any)["title"] = "changed details"
+
+	second := prepared.Details()
+	if second.Messages[0].Content != "Input=original artifact Label=original variable" {
+		t.Fatalf("details message changed: %q", second.Messages[0].Content)
+	}
+	if second.InputHashes["input"] != "original-input-hash" {
+		t.Fatalf("details input hash changed: %q", second.InputHashes["input"])
+	}
+	if second.EffectiveModelParams.Model != "original-model" ||
+		second.EffectiveModelParams.Temperature != 0.25 ||
+		second.EffectiveModelParams.ExtraParams["nested"].(map[string]any)["source"] != "original" {
+		t.Fatalf("details target changed: %+v", second.EffectiveModelParams)
+	}
+	schema := second.StructuredOutput.JSONSchema.Schema.(map[string]any)
+	if schema["title"] != "original root" {
+		t.Fatalf("details schema changed: %#v", schema)
+	}
+
+	result, err := engine.RunPrepared(context.Background(), prepared)
+	if err != nil {
+		t.Fatalf("run prepared after preparation-context cancellation: %v", err)
+	}
+	if result.Validation.Status != promptkit.ValidationPassed || !result.Validation.IsValid {
+		t.Fatalf("frozen schema did not validate original output: %+v", result.Validation)
+	}
+	if reader.callCount() != 1 {
+		t.Fatalf("execution reopened artifact source: calls=%d", reader.callCount())
+	}
+
+	requests := client.snapshot()
+	if len(requests) != 1 {
+		t.Fatalf("generation calls=%d, want 1", len(requests))
+	}
+	generated := requests[0]
+	if generated.Prompt.Messages[0].Content != second.Messages[0].Content ||
+		generated.Target.Model != second.EffectiveModelParams.Model ||
+		!reflect.DeepEqual(generated.Target.ExtraParams, second.EffectiveModelParams.ExtraParams) ||
+		!reflect.DeepEqual(generated.StructuredOutput, second.StructuredOutput) {
+		t.Fatalf("generation did not use frozen details:\nrequest=%+v\ndetails=%+v", generated, second)
+	}
+	if result.PromptID != second.PromptID ||
+		result.PromptVersion != second.PromptVersion ||
+		result.PromptHash != second.PromptHash ||
+		result.SessionID != second.SessionID ||
+		result.RenderedPromptHash != second.RenderedPromptHash ||
+		result.SelectedProfileID != second.SelectedProfileID ||
+		result.SelectedBackendID != second.SelectedBackendID ||
+		!reflect.DeepEqual(result.EffectiveModelParams, second.EffectiveModelParams) ||
+		!reflect.DeepEqual(result.InputHashes, second.InputHashes) {
+		t.Fatalf("result provenance does not match details:\nresult=%+v\ndetails=%+v", result, second)
+	}
+}
+
+func TestPreparedExecutionLifecycleAndEngineBinding(t *testing.T) {
+	ownerClient := &preparedRecordingClient{
+		response: &promptkit.GenerateResponse{Content: "ok"},
+	}
+	owner := newPreparedContractEngine(t, ownerClient, "owner content")
+	foreign := newPreparedContractEngine(t, &preparedRecordingClient{
+		response: &promptkit.GenerateResponse{Content: "unexpected"},
+	}, "foreign content")
+
+	prepared, err := owner.PrepareExecution(context.Background(), promptkit.RunRequest{PromptID: "prepared"})
+	if err != nil {
+		t.Fatalf("prepare execution: %v", err)
+	}
+	copied := *prepared
+
+	var nilEngine *promptkit.Engine
+	if result, err := nilEngine.RunPrepared(context.Background(), prepared); result != nil ||
+		!errors.Is(err, promptkit.ErrInvalidConfig) {
+		t.Fatalf("nil engine result=(%+v, %v), want ErrInvalidConfig", result, err)
+	}
+	if result, err := foreign.RunPrepared(context.Background(), prepared); result != nil ||
+		!errors.Is(err, promptkit.ErrInvalidRequest) {
+		t.Fatalf("foreign engine result=(%+v, %v), want ErrInvalidRequest", result, err)
+	}
+	if result, err := owner.RunPrepared(context.Background(), nil); result != nil ||
+		!errors.Is(err, promptkit.ErrInvalidRequest) {
+		t.Fatalf("nil handle result=(%+v, %v), want ErrInvalidRequest", result, err)
+	}
+	if result, err := owner.RunPrepared(context.Background(), &promptkit.PreparedExecution{}); result != nil ||
+		!errors.Is(err, promptkit.ErrInvalidRequest) {
+		t.Fatalf("zero handle result=(%+v, %v), want ErrInvalidRequest", result, err)
+	}
+
+	result, err := owner.RunPrepared(context.Background(), &copied)
+	if err != nil || result == nil {
+		t.Fatalf("owner run prepared=(%+v, %v), want success", result, err)
+	}
+	for name, handle := range map[string]*promptkit.PreparedExecution{
+		"original": prepared,
+		"copy":     &copied,
+	} {
+		if result, err := owner.RunPrepared(context.Background(), handle); result != nil ||
+			!errors.Is(err, promptkit.ErrInvalidRequest) {
+			t.Fatalf("%s reused handle result=(%+v, %v), want ErrInvalidRequest", name, result, err)
+		}
+		if handle.Details().PromptID != "prepared" {
+			t.Fatalf("%s details unavailable after execution", name)
+		}
+	}
+	if len(ownerClient.snapshot()) != 1 {
+		t.Fatalf("owner generation calls=%d, want 1", len(ownerClient.snapshot()))
+	}
+
+	collaboratorFailure := errors.New("prepared collaborator failure")
+	failingClient := &preparedRecordingClient{err: collaboratorFailure}
+	failingEngine := newPreparedContractEngine(t, failingClient, "failure content")
+	failing, err := failingEngine.PrepareExecution(context.Background(), promptkit.RunRequest{PromptID: "prepared"})
+	if err != nil {
+		t.Fatalf("prepare failing execution: %v", err)
+	}
+	if result, err := failingEngine.RunPrepared(context.Background(), failing); result != nil ||
+		!errors.Is(err, promptkit.ErrLLMGenerate) ||
+		!errors.Is(err, collaboratorFailure) {
+		t.Fatalf("generation failure result=(%+v, %v), want public and collaborator identities", result, err)
+	}
+	if result, err := failingEngine.RunPrepared(context.Background(), failing); result != nil ||
+		!errors.Is(err, promptkit.ErrInvalidRequest) {
+		t.Fatalf("failed execution was reusable: result=(%+v, %v)", result, err)
+	}
+
+	cancellationRelease := make(chan struct{})
+	cancellationStarted := make(chan struct{}, 1)
+	cancelingEngine := newPreparedContractEngine(t, &preparedRecordingClient{
+		response: &promptkit.GenerateResponse{Content: "unexpected"},
+		started:  cancellationStarted,
+		release:  cancellationRelease,
+	}, "cancellation content")
+	canceling, err := cancelingEngine.PrepareExecution(
+		context.Background(),
+		promptkit.RunRequest{PromptID: "prepared"},
+	)
+	if err != nil {
+		t.Fatalf("prepare canceled execution: %v", err)
+	}
+	executionContext, cancelExecution := context.WithCancel(context.Background())
+	type canceledOutcome struct {
+		result *promptkit.RunResult
+		err    error
+	}
+	canceledResult := make(chan canceledOutcome, 1)
+	go func() {
+		result, runErr := cancelingEngine.RunPrepared(executionContext, canceling)
+		canceledResult <- canceledOutcome{result: result, err: runErr}
+	}()
+	select {
+	case <-cancellationStarted:
+	case <-time.After(5 * time.Second):
+		t.Fatal("timed out waiting for cancelable generation")
+	}
+	cancelExecution()
+	select {
+	case outcome := <-canceledResult:
+		if outcome.result != nil ||
+			!errors.Is(outcome.err, promptkit.ErrLLMGenerate) ||
+			!errors.Is(outcome.err, context.Canceled) {
+			t.Fatalf(
+				"canceled execution=(%+v, %v), want generation and context identities",
+				outcome.result,
+				outcome.err,
+			)
+		}
+	case <-time.After(5 * time.Second):
+		t.Fatal("timed out waiting for canceled execution")
+	}
+	if result, err := cancelingEngine.RunPrepared(context.Background(), canceling); result != nil ||
+		!errors.Is(err, promptkit.ErrInvalidRequest) {
+		t.Fatalf("canceled execution was reusable: result=(%+v, %v)", result, err)
+	}
+}
+
+func TestPreparedExecutionConcurrentClaimAllowsOneGeneration(t *testing.T) {
+	release := make(chan struct{})
+	client := &preparedRecordingClient{
+		response: &promptkit.GenerateResponse{Content: "ok"},
+		started:  make(chan struct{}, 1),
+		release:  release,
+	}
+	engine := newPreparedContractEngine(t, client, "concurrent content")
+	prepared, err := engine.PrepareExecution(context.Background(), promptkit.RunRequest{PromptID: "prepared"})
+	if err != nil {
+		t.Fatalf("prepare execution: %v", err)
+	}
+
+	type outcome struct {
+		result *promptkit.RunResult
+		err    error
+	}
+	outcomes := make(chan outcome, 2)
+	for i := 0; i < 2; i++ {
+		go func() {
+			result, runErr := engine.RunPrepared(context.Background(), prepared)
+			outcomes <- outcome{result: result, err: runErr}
+		}()
+	}
+
+	select {
+	case <-client.started:
+	case <-time.After(5 * time.Second):
+		t.Fatal("timed out waiting for generation")
+	}
+	select {
+	case loser := <-outcomes:
+		if loser.result != nil || !errors.Is(loser.err, promptkit.ErrInvalidRequest) {
+			t.Fatalf("concurrent loser=(%+v, %v), want ErrInvalidRequest", loser.result, loser.err)
+		}
+	case <-time.After(5 * time.Second):
+		t.Fatal("timed out waiting for rejected concurrent claim")
+	}
+
+	close(release)
+	select {
+	case winner := <-outcomes:
+		if winner.err != nil || winner.result == nil {
+			t.Fatalf("concurrent winner=(%+v, %v), want success", winner.result, winner.err)
+		}
+	case <-time.After(5 * time.Second):
+		t.Fatal("timed out waiting for successful concurrent claim")
+	}
+	if len(client.snapshot()) != 1 {
+		t.Fatalf("generation calls=%d, want 1", len(client.snapshot()))
+	}
+}
+
+func TestPreparedExecutionRunAndDiscardRaceHasOneWinner(t *testing.T) {
+	const attempts = 32
+
+	for i := 0; i < attempts; i++ {
+		client := &preparedRecordingClient{
+			response: &promptkit.GenerateResponse{Content: "ok"},
+		}
+		engine := newPreparedContractEngine(t, client, "race content")
+		prepared, err := engine.PrepareExecution(
+			context.Background(),
+			promptkit.RunRequest{PromptID: "prepared"},
+		)
+		if err != nil {
+			t.Fatalf("attempt %d prepare execution: %v", i, err)
+		}
+
+		start := make(chan struct{})
+		type outcome struct {
+			result *promptkit.RunResult
+			err    error
+		}
+		runOutcome := make(chan outcome, 1)
+		discardDone := make(chan struct{})
+
+		go func() {
+			<-start
+			result, runErr := engine.RunPrepared(context.Background(), prepared)
+			runOutcome <- outcome{result: result, err: runErr}
+		}()
+		go func() {
+			<-start
+			prepared.Discard()
+			close(discardDone)
+		}()
+
+		close(start)
+		runResult := <-runOutcome
+		<-discardDone
+		calls := len(client.snapshot())
+		switch {
+		case runResult.err == nil:
+			if runResult.result == nil || calls != 1 {
+				t.Fatalf(
+					"attempt %d run won with outcome=(%+v, %v), generation calls=%d",
+					i,
+					runResult.result,
+					runResult.err,
+					calls,
+				)
+			}
+		case errors.Is(runResult.err, promptkit.ErrInvalidRequest):
+			if runResult.result != nil || calls != 0 {
+				t.Fatalf(
+					"attempt %d discard won with outcome=(%+v, %v), generation calls=%d",
+					i,
+					runResult.result,
+					runResult.err,
+					calls,
+				)
+			}
+		default:
+			t.Fatalf("attempt %d unexpected run outcome=(%+v, %v)", i, runResult.result, runResult.err)
+		}
+		if prepared.Details().PromptID != "prepared" {
+			t.Fatalf("attempt %d details unavailable after race", i)
+		}
+	}
+}
+
+func TestPreparedExecutionDiscardAndFormattingDoNotExposePrivateState(t *testing.T) {
+	const (
+		directCredential = "pk-test-direct-credential-41f7"
+		renderedContent  = "rendered-content-sentinel-98d2"
+	)
+	client := &preparedRecordingClient{
+		response: &promptkit.GenerateResponse{Content: "generated output"},
+	}
+	engine, err := promptkit.NewEngine(
+		promptkit.Config{},
+		promptkit.WithPromptFS(contractPromptFS("prepared", "profile", renderedContent), "."),
+		promptkit.WithProfiles(promptkit.Profile{
+			ID:             "profile",
+			Endpoint:       "http://example.test/v1",
+			Model:          "model",
+			APIKeyRequired: true,
+		}),
+		promptkit.WithLLMClient(client),
+	)
+	if err != nil {
+		t.Fatalf("construct engine: %v", err)
+	}
+	prepared, err := engine.PrepareExecution(context.Background(), promptkit.RunRequest{
+		PromptID: "prepared",
+		APIKey:   directCredential,
+	})
+	if err != nil {
+		t.Fatalf("prepare execution: %v", err)
+	}
+
+	formattedValues := []string{
+		fmt.Sprint(prepared),
+		fmt.Sprintf("%+v", prepared),
+		fmt.Sprintf("%#v", prepared),
+	}
+	for _, formatted := range formattedValues {
+		if formatted != "promptkit.PreparedExecution{opaque}" {
+			t.Fatalf("unexpected opaque formatting: %q", formatted)
+		}
+		assertPreparedPrivateValuesAbsent(t, formatted, directCredential, renderedContent)
+	}
+	payload, err := json.Marshal(prepared)
+	if err != nil {
+		t.Fatalf("marshal opaque handle: %v", err)
+	}
+	assertPreparedPrivateValuesAbsent(t, string(payload), directCredential, renderedContent)
+
+	detailsBefore := prepared.Details()
+	detailsJSON, err := json.Marshal(detailsBefore)
+	if err != nil {
+		t.Fatalf("marshal prepared details: %v", err)
+	}
+	assertPreparedPrivateValuesAbsent(t, string(detailsJSON), directCredential)
+
+	prepared.Discard()
+	prepared.Discard()
+	result, lifecycleErr := engine.RunPrepared(context.Background(), prepared)
+	if result != nil || !errors.Is(lifecycleErr, promptkit.ErrInvalidRequest) {
+		t.Fatalf("discarded execution result=(%+v, %v), want ErrInvalidRequest", result, lifecycleErr)
+	}
+	assertPreparedPrivateValuesAbsent(t, lifecycleErr.Error(), directCredential, renderedContent)
+	if !reflect.DeepEqual(prepared.Details(), detailsBefore) {
+		t.Fatal("details changed after discard")
+	}
+
+	executed, err := engine.PrepareExecution(context.Background(), promptkit.RunRequest{
+		PromptID: "prepared",
+		APIKey:   directCredential,
+	})
+	if err != nil {
+		t.Fatalf("prepare execution for request inspection: %v", err)
+	}
+	executionResult, err := engine.RunPrepared(context.Background(), executed)
+	if err != nil {
+		t.Fatalf("run execution for request inspection: %v", err)
+	}
+	requests := client.snapshot()
+	if len(requests) != 1 || requests[0].APIKey != directCredential {
+		t.Fatalf("direct credential did not reach only the client credential field: %#v", requests)
+	}
+	requestJSON, err := json.Marshal(requests[0])
+	if err != nil {
+		t.Fatalf("marshal captured generate request: %v", err)
+	}
+	for _, value := range []string{
+		fmt.Sprint(requests[0]),
+		fmt.Sprintf("%+v", requests[0]),
+		fmt.Sprintf("%#v", requests[0]),
+		string(requestJSON),
+		fmt.Sprint(executionResult),
+	} {
+		assertPreparedPrivateValuesAbsent(t, value, directCredential)
+	}
+	resultJSON, err := json.Marshal(executionResult)
+	if err != nil {
+		t.Fatalf("marshal execution result: %v", err)
+	}
+	assertPreparedPrivateValuesAbsent(t, string(resultJSON), directCredential)
+
+	var nilHandle *promptkit.PreparedExecution
+	nilHandle.Discard()
+	if !reflect.DeepEqual(nilHandle.Details(), promptkit.PreparedRun{}) {
+		t.Fatalf("nil handle details=%+v, want zero value", nilHandle.Details())
+	}
+	zeroHandle := &promptkit.PreparedExecution{}
+	zeroHandle.Discard()
+	if !reflect.DeepEqual(zeroHandle.Details(), promptkit.PreparedRun{}) {
+		t.Fatalf("zero handle details=%+v, want zero value", zeroHandle.Details())
+	}
+}
+
+func TestPreparedExecutionCredentialCapacityAndTimingBoundaries(t *testing.T) {
+	t.Run("credential is rechecked before generation", func(t *testing.T) {
+		const (
+			environmentName = "PROMPTKIT_PREPARED_CONTRACT_KEY"
+			environmentKey  = "environment-credential-sentinel"
+		)
+		t.Setenv(environmentName, environmentKey)
+		client := &preparedRecordingClient{
+			response: &promptkit.GenerateResponse{Content: "unexpected"},
+		}
+		engine, err := promptkit.NewEngine(
+			promptkit.Config{},
+			promptkit.WithPromptFS(contractPromptFS("prepared", "profile", "content"), "."),
+			promptkit.WithProfileFS(preparedCredentialProfileSource(environmentName), "."),
+			promptkit.WithLLMClient(client),
+		)
+		if err != nil {
+			t.Fatalf("construct credential engine: %v", err)
+		}
+		prepared, err := engine.PrepareExecution(context.Background(), promptkit.RunRequest{PromptID: "prepared"})
+		if err != nil {
+			t.Fatalf("prepare credential execution: %v", err)
+		}
+		if err := os.Unsetenv(environmentName); err != nil {
+			t.Fatalf("unset credential environment: %v", err)
+		}
+
+		result, err := engine.RunPrepared(context.Background(), prepared)
+		if result != nil ||
+			!errors.Is(err, promptkit.ErrInvalidRequest) ||
+			!errors.Is(err, promptkit.ErrAPIKeyEnvMissing) {
+			t.Fatalf("credential execution=(%+v, %v), want credential identities", result, err)
+		}
+		if len(client.snapshot()) != 0 {
+			t.Fatalf("credential failure reached generation: %d calls", len(client.snapshot()))
+		}
+		assertPreparedPrivateValuesAbsent(t, err.Error(), environmentKey)
+		if result, err := engine.RunPrepared(context.Background(), prepared); result != nil ||
+			!errors.Is(err, promptkit.ErrInvalidRequest) {
+			t.Fatalf("credential failure did not consume handle: result=(%+v, %v)", result, err)
+		}
+	})
+
+	t.Run("preparation does not admit and execution timing starts after retention", func(t *testing.T) {
+		release := make(chan struct{})
+		client := newCapacityGateClient(release, 4)
+		engine := newBackendCapacityEngine(t, client, 1, capacityInt(0), nil)
+
+		activeRun := make(chan capacityRunResult, 1)
+		go runCapacityRequest(
+			engine,
+			context.Background(),
+			promptkit.RunRequest{PromptID: "prompt"},
+			activeRun,
+		)
+		awaitCapacityRequest(t, client.started)
+
+		prepared, err := engine.PrepareExecution(
+			context.Background(),
+			promptkit.RunRequest{PromptID: "prompt"},
+		)
+		if err != nil {
+			t.Fatalf("prepare while capacity is full: %v", err)
+		}
+		if _, _, calls := client.snapshot(); calls != 1 {
+			t.Fatalf("preparation invoked generation: calls=%d", calls)
+		}
+		if result, err := engine.RunPrepared(context.Background(), prepared); result != nil ||
+			!errors.Is(err, promptkit.ErrCapacityExceeded) {
+			t.Fatalf("capacity execution=(%+v, %v), want ErrCapacityExceeded", result, err)
+		} else {
+			var capacityErr *promptkit.CapacityError
+			if !errors.As(err, &capacityErr) || capacityErr == nil || capacityErr.BackendID != "limited" {
+				t.Fatalf("capacity execution=%v, want limited CapacityError", err)
+			}
+		}
+		if result, err := engine.RunPrepared(context.Background(), prepared); result != nil ||
+			!errors.Is(err, promptkit.ErrInvalidRequest) {
+			t.Fatalf("capacity rejection did not consume handle: result=(%+v, %v)", result, err)
+		}
+		if prepared.Details().PromptID != "prompt" {
+			t.Fatal("details unavailable after capacity rejection")
+		}
+
+		close(release)
+		activeOutcome := awaitCapacityRun(t, activeRun)
+		if activeOutcome.err != nil || activeOutcome.result == nil {
+			t.Fatalf("active run outcome=(%+v, %v), want success", activeOutcome.result, activeOutcome.err)
+		}
+
+		timed, err := engine.PrepareExecution(
+			context.Background(),
+			promptkit.RunRequest{PromptID: "prompt"},
+		)
+		if err != nil {
+			t.Fatalf("prepare timed execution: %v", err)
+		}
+		details := timed.Details()
+		time.Sleep(25 * time.Millisecond)
+		executionFloor := time.Now().UTC()
+		result, err := engine.RunPrepared(context.Background(), timed)
+		if err != nil {
+			t.Fatalf("run timed execution: %v", err)
+		}
+		if result.StartTime.Before(executionFloor) ||
+			!result.StartTime.After(details.EndTime) ||
+			result.EndTime.Before(result.StartTime) ||
+			result.Duration != result.EndTime.Sub(result.StartTime) {
+			t.Fatalf(
+				"execution timing includes preparation or retention: details_end=%s floor=%s result=%+v",
+				details.EndTime,
+				executionFloor,
+				result,
+			)
+		}
+		if _, _, calls := client.snapshot(); calls != 2 {
+			t.Fatalf("generation calls=%d, want active and timed executions only", calls)
+		}
+	})
+}
+
+type mutablePreparedArtifactReader struct {
+	mu    sync.Mutex
+	body  string
+	hash  string
+	calls int
+}
+
+func (r *mutablePreparedArtifactReader) Read(
+	_ context.Context,
+	_ promptkit.ArtifactRef,
+) (*promptkit.Artifact, error) {
+	r.mu.Lock()
+	defer r.mu.Unlock()
+	r.calls++
+	return &promptkit.Artifact{
+		Body: []byte(r.body),
+		Hash: r.hash,
+	}, nil
+}
+
+func (r *mutablePreparedArtifactReader) set(body, hash string) {
+	r.mu.Lock()
+	defer r.mu.Unlock()
+	r.body = body
+	r.hash = hash
+}
+
+func (r *mutablePreparedArtifactReader) callCount() int {
+	r.mu.Lock()
+	defer r.mu.Unlock()
+	return r.calls
+}
+
+type preparedRecordingClient struct {
+	mu       sync.Mutex
+	response *promptkit.GenerateResponse
+	err      error
+	requests []promptkit.GenerateRequest
+	started  chan struct{}
+	release  <-chan struct{}
+}
+
+func (c *preparedRecordingClient) Generate(
+	ctx context.Context,
+	request promptkit.GenerateRequest,
+) (*promptkit.GenerateResponse, error) {
+	c.mu.Lock()
+	c.requests = append(c.requests, request)
+	c.mu.Unlock()
+
+	if c.started != nil {
+		c.started <- struct{}{}
+	}
+	if c.release != nil {
+		select {
+		case <-c.release:
+		case <-ctx.Done():
+			return nil, ctx.Err()
+		}
+	}
+	if c.err != nil {
+		return nil, c.err
+	}
+	return c.response, nil
+}
+
+func (c *preparedRecordingClient) snapshot() []promptkit.GenerateRequest {
+	c.mu.Lock()
+	defer c.mu.Unlock()
+	return append([]promptkit.GenerateRequest(nil), c.requests...)
+}
+
+func newPreparedContractEngine(
+	t *testing.T,
+	client promptkit.LLMClient,
+	message string,
+) *promptkit.Engine {
+	t.Helper()
+	engine, err := promptkit.NewEngine(
+		promptkit.Config{},
+		promptkit.WithPromptFS(contractPromptFS("prepared", "profile", message), "."),
+		promptkit.WithProfiles(promptkit.Profile{
+			ID:       "profile",
+			Endpoint: "http://example.test/v1",
+			Model:    "model",
+		}),
+		promptkit.WithLLMClient(client),
+	)
+	if err != nil {
+		t.Fatalf("construct prepared execution engine: %v", err)
+	}
+	return engine
+}
+
+func preparedPromptSource(label string) fstest.MapFS {
+	return fstest.MapFS{
+		"prompt.yaml": &fstest.MapFile{Data: []byte(`id: prepared
+version: "1"
+default_profile: profile
+inputs:
+  - name: input
+    required: true
+messages:
+  - role: user
+    content: 'Input={{input "input"}} Label={{.label}}'
+description: ` + label + `
+output:
+  format: json
+  validation_mode: json_schema
+  schema_path: schema.json
+`)},
+	}
+}
+
+func preparedProfileSource(model string) fstest.MapFS {
+	return fstest.MapFS{
+		"profile.yaml": &fstest.MapFile{Data: []byte(`id: profile
+endpoint: http://example.test/v1
+model: ` + model + `
+`)},
+	}
+}
+
+func preparedCredentialProfileSource(environmentName string) fstest.MapFS {
+	return fstest.MapFS{
+		"profile.yaml": &fstest.MapFile{Data: []byte(`id: profile
+endpoint: http://example.test/v1
+model: model
+api_key_env: ` + environmentName + `
+`)},
+	}
+}
+
+func preparedSchemaSource() fstest.MapFS {
+	return fstest.MapFS{
+		"schema.json": &fstest.MapFile{Data: []byte(`{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "title": "original root",
+  "type": "object",
+  "required": ["value"],
+  "properties": {
+    "value": {"$ref": "value.json"}
+  }
+}`)},
+		"value.json": &fstest.MapFile{Data: []byte(`{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "type": "integer",
+  "minimum": 2
+}`)},
+	}
+}
+
+func assertPreparedPrivateValuesAbsent(t *testing.T, value string, privateValues ...string) {
+	t.Helper()
+	for _, privateValue := range privateValues {
+		if strings.Contains(value, privateValue) {
+			t.Fatalf("value exposed private data %q: %s", privateValue, value)
+		}
+	}
+}

public_contract_test.go

@@ -5,6 +5,8 @@ import (
 	"encoding/json"
 	"errors"
 	"fmt"
+	"io/fs"
+	"reflect"
 	"strings"
 	"sync"
 	"sync/atomic"
@@ -15,6 +17,380 @@ import (
 	"gitea.maximumdirect.net/eric/promptkit"
 )
 
+type inspectionCountingFS struct {
+	opens atomic.Int64
+}
+
+func (f *inspectionCountingFS) Open(string) (fs.File, error) {
+	f.opens.Add(1)
+	return nil, fs.ErrNotExist
+}
+
+func TestInspectProfileResolvesCredentialStatesWithoutPromptOrGeneration(t *testing.T) {
+	const environmentName = "PROMPTKIT_INSPECTION_ABSENT_KEY"
+	t.Setenv(environmentName, "")
+	client := &fakeLLMClient{response: &promptkit.GenerateResponse{Content: "unexpected"}}
+	engine, err := promptkit.NewEngine(promptkit.Config{},
+		promptkit.WithPromptFS(fstest.MapFS{}, "."),
+		promptkit.WithProfileFS(fstest.MapFS{
+			"environment.yaml": &fstest.MapFile{Data: []byte(`id: environment
+endpoint: http://environment.example/v1
+model: environment-model
+api_key_env: PROMPTKIT_INSPECTION_ABSENT_KEY
+`)},
+		}, "."),
+		promptkit.WithProfiles(
+			promptkit.Profile{ID: "direct", Endpoint: "http://direct.example/v1", Model: "direct-model", APIKeyRequired: true},
+			promptkit.Profile{ID: "none", Endpoint: "http://none.example/v1", Model: "none-model"},
+		),
+		promptkit.WithLLMClient(client),
+	)
+	if err != nil {
+		t.Fatalf("construct inspection engine: %v", err)
+	}
+
+	for _, tc := range []struct {
+		profileID     string
+		wantEnv       string
+		wantDirectKey bool
+		wantEndpoint  string
+	}{
+		{profileID: " environment ", wantEnv: environmentName, wantEndpoint: "http://environment.example/v1"},
+		{profileID: "direct", wantDirectKey: true, wantEndpoint: "http://direct.example/v1"},
+		{profileID: "none", wantEndpoint: "http://none.example/v1"},
+	} {
+		t.Run(tc.profileID, func(t *testing.T) {
+			inspection, err := engine.InspectProfile(context.Background(), tc.profileID)
+			if err != nil {
+				t.Fatalf("inspect profile: %v", err)
+			}
+			if inspection.ProfileID != strings.TrimSpace(tc.profileID) ||
+				inspection.EffectiveModelParams.Endpoint != tc.wantEndpoint ||
+				inspection.EffectiveModelParams.BackendID != "" ||
+				inspection.EffectiveModelParams.APIKeyEnv != tc.wantEnv ||
+				inspection.APIKeyRequired != tc.wantDirectKey {
+				t.Fatalf("unexpected inspection: %#v", inspection)
+			}
+		})
+	}
+	if len(client.requests) != 0 {
+		t.Fatalf("inspection invoked the model client %d times", len(client.requests))
+	}
+}
+
+func TestInspectProfilePreservesPublicErrorIdentities(t *testing.T) {
+	newEngine := func(t *testing.T, options ...promptkit.Option) *promptkit.Engine {
+		t.Helper()
+		engine, err := promptkit.NewEngine(
+			promptkit.Config{},
+			append([]promptkit.Option{promptkit.WithPromptFS(fstest.MapFS{}, ".")}, options...)...,
+		)
+		if err != nil {
+			t.Fatalf("construct inspection engine: %v", err)
+		}
+		return engine
+	}
+
+	var nilEngine *promptkit.Engine
+	if result, err := nilEngine.InspectProfile(context.Background(), "profile"); result != nil ||
+		!errors.Is(err, promptkit.ErrInvalidConfig) {
+		t.Fatalf("nil engine result=(%#v, %v), want ErrInvalidConfig", result, err)
+	}
+
+	valid := newEngine(t, promptkit.WithProfiles(promptkit.Profile{
+		ID: "profile", Endpoint: "http://profile.example/v1", Model: "model",
+	}))
+	if result, err := valid.InspectProfile(context.Background(), " \t "); result != nil ||
+		!errors.Is(err, promptkit.ErrInvalidRequest) {
+		t.Fatalf("blank profile result=(%#v, %v), want ErrInvalidRequest", result, err)
+	}
+	if result, err := valid.InspectProfile(context.Background(), "missing"); result != nil ||
+		!errors.Is(err, promptkit.ErrProfileNotFound) || errors.Is(err, promptkit.ErrProfileLoad) {
+		t.Fatalf("missing profile result=(%#v, %v), want only ErrProfileNotFound", result, err)
+	}
+
+	malformed := newEngine(t, promptkit.WithProfileFS(fstest.MapFS{
+		"broken.yaml": &fstest.MapFile{Data: []byte("id: broken\nendpoint: http://broken.example/v1\nmodel: model\nunknown: value\n")},
+	}, "."))
+	if result, err := malformed.InspectProfile(context.Background(), "broken"); result != nil ||
+		!errors.Is(err, promptkit.ErrProfileLoad) {
+		t.Fatalf("malformed profile result=(%#v, %v), want ErrProfileLoad", result, err)
+	}
+
+	unknownBackend := newEngine(t, promptkit.WithProfiles(promptkit.Profile{
+		ID: "unknown-backend", BackendID: "unknown", Model: "model",
+	}))
+	if result, err := unknownBackend.InspectProfile(context.Background(), "unknown-backend"); result != nil ||
+		!errors.Is(err, promptkit.ErrProfileLoad) {
+		t.Fatalf("unknown backend result=(%#v, %v), want ErrProfileLoad", result, err)
+	}
+
+	countingFS := &inspectionCountingFS{}
+	canceled := newEngine(t, promptkit.WithProfileFS(countingFS, "."))
+	ctx, cancel := context.WithCancel(context.Background())
+	cancel()
+	if result, err := canceled.InspectProfile(ctx, "profile"); result != nil ||
+		!errors.Is(err, promptkit.ErrProfileLoad) || !errors.Is(err, context.Canceled) || countingFS.opens.Load() != 0 {
+		t.Fatalf("canceled inspection result=(%#v, %v), opens=%d", result, err, countingFS.opens.Load())
+	}
+}
+
+func TestInspectProfileReturnsIndependentTargetMatchingPreparation(t *testing.T) {
+	extraParams := map[string]any{
+		"nested": map[string]any{"value": "original"},
+	}
+	engine, err := promptkit.NewEngine(promptkit.Config{},
+		promptkit.WithPromptFS(contractPromptFS("prompt", "profile", "message"), "."),
+		promptkit.WithProfiles(promptkit.Profile{
+			ID: "profile", Endpoint: "http://profile.example/v1", Model: "model", ExtraParams: extraParams,
+		}),
+	)
+	if err != nil {
+		t.Fatalf("construct inspection engine: %v", err)
+	}
+
+	first, err := engine.InspectProfile(context.Background(), "profile")
+	if err != nil {
+		t.Fatalf("first inspection: %v", err)
+	}
+	first.EffectiveModelParams.ExtraParams["nested"].(map[string]any)["value"] = "changed"
+	first.EffectiveModelParams.ExtraParams["later"] = true
+
+	second, err := engine.InspectProfile(context.Background(), "profile")
+	if err != nil {
+		t.Fatalf("second inspection: %v", err)
+	}
+	prepared, err := engine.Prepare(context.Background(), promptkit.RunRequest{PromptID: "prompt"})
+	if err != nil {
+		t.Fatalf("prepare after inspection mutation: %v", err)
+	}
+	for _, target := range []promptkit.ExecutionTarget{second.EffectiveModelParams, prepared.EffectiveModelParams} {
+		if target.ExtraParams["nested"].(map[string]any)["value"] != "original" || target.ExtraParams["later"] != nil {
+			t.Fatalf("inspection mutation reached engine-owned target: %#v", target.ExtraParams)
+		}
+	}
+	if !reflect.DeepEqual(second.EffectiveModelParams, prepared.EffectiveModelParams) {
+		t.Fatalf("inspection target=%#v, preparation target=%#v", second.EffectiveModelParams, prepared.EffectiveModelParams)
+	}
+}
+
+func TestInspectPromptReturnsDeclaredMetadataWithoutExecutionWork(t *testing.T) {
+	client := &fakeLLMClient{response: &promptkit.GenerateResponse{Content: "unexpected"}}
+	engine, err := promptkit.NewEngine(promptkit.Config{},
+		promptkit.WithPromptFS(fstest.MapFS{
+			"report-v1.yaml": &fstest.MapFile{Data: []byte(`id: report
+version: "1.0.0"
+messages:
+  - role: user
+    content: old report
+output:
+  format: text
+  validation_mode: none
+`)},
+			"report-v2.yaml": &fstest.MapFile{Data: []byte(`id: report
+version: "2.0.0"
+default_profile: missing-profile
+inputs:
+  - name: location
+    required: true
+    content_type: text/plain
+    description: Forecast location.
+  - name: units
+    content_type: text/plain
+    description: Unit preference.
+messages:
+  - role: user
+    content_file: messages/report.md
+output:
+  format: json
+  validation_mode: json_schema
+  schema_path: schemas/report.json
+`)},
+			"messages/report.md": &fstest.MapFile{Data: []byte("rendered report body is not returned")},
+		}, "."),
+		promptkit.WithLLMClient(client),
+	)
+	if err != nil {
+		t.Fatalf("construct prompt-inspection engine: %v", err)
+	}
+
+	inspection, err := engine.InspectPrompt(context.Background(), "report", "2.0.0")
+	if err != nil {
+		t.Fatalf("inspect prompt: %v", err)
+	}
+	if inspection.PromptID != "report" ||
+		inspection.PromptVersion != "2.0.0" ||
+		inspection.PromptHash == "" ||
+		inspection.DefaultProfileID != "missing-profile" ||
+		inspection.OutputContract != (promptkit.OutputContract{
+			Format:         promptkit.FormatJSON,
+			ValidationMode: promptkit.ValidationJSONSchema,
+			SchemaPath:     "schemas/report.json",
+		}) {
+		t.Fatalf("unexpected inspection metadata: %#v", inspection)
+	}
+	wantInputs := []promptkit.PromptInputDefinition{
+		{Name: "location", Required: true, ContentType: "text/plain", Description: "Forecast location."},
+		{Name: "units", ContentType: "text/plain", Description: "Unit preference."},
+	}
+	if !reflect.DeepEqual(inspection.Inputs, wantInputs) {
+		t.Fatalf("inspection inputs=%#v, want %#v", inspection.Inputs, wantInputs)
+	}
+	if len(client.requests) != 0 {
+		t.Fatalf("inspection invoked the model client %d times", len(client.requests))
+	}
+}
+
+func TestInspectPromptPreservesPublicErrorIdentities(t *testing.T) {
+	newEngine := func(t *testing.T, source fstest.MapFS) *promptkit.Engine {
+		t.Helper()
+		engine, err := promptkit.NewEngine(promptkit.Config{}, promptkit.WithPromptFS(source, "."))
+		if err != nil {
+			t.Fatalf("construct prompt-inspection engine: %v", err)
+		}
+		return engine
+	}
+	validSource := fstest.MapFS{
+		"prompt.yaml": &fstest.MapFile{Data: []byte(`id: prompt
+version: "1"
+messages:
+  - role: user
+    content: body
+output:
+  format: text
+  validation_mode: none
+`)},
+	}
+
+	var nilEngine *promptkit.Engine
+	if result, err := nilEngine.InspectPrompt(context.Background(), "prompt", "1"); result != nil ||
+		!errors.Is(err, promptkit.ErrInvalidConfig) {
+		t.Fatalf("nil engine result=(%#v, %v), want ErrInvalidConfig", result, err)
+	}
+
+	valid := newEngine(t, validSource)
+	if result, err := valid.InspectPrompt(context.Background(), " \t ", "1"); result != nil ||
+		!errors.Is(err, promptkit.ErrInvalidRequest) {
+		t.Fatalf("blank prompt result=(%#v, %v), want ErrInvalidRequest", result, err)
+	}
+	if result, err := valid.InspectPrompt(context.Background(), "missing", "1"); result != nil ||
+		!errors.Is(err, promptkit.ErrPromptNotFound) || errors.Is(err, promptkit.ErrPromptLoad) {
+		t.Fatalf("missing prompt result=(%#v, %v), want only ErrPromptNotFound", result, err)
+	}
+	if result, err := valid.InspectPrompt(context.Background(), "prompt", "missing"); result != nil ||
+		!errors.Is(err, promptkit.ErrPromptNotFound) || errors.Is(err, promptkit.ErrPromptLoad) {
+		t.Fatalf("missing version result=(%#v, %v), want only ErrPromptNotFound", result, err)
+	}
+
+	ambiguous := newEngine(t, fstest.MapFS{
+		"one.yaml": &fstest.MapFile{Data: []byte(`id: prompt
+version: "1"
+messages:
+  - role: user
+    content: first
+output:
+  format: text
+  validation_mode: none
+`)},
+		"two.yaml": &fstest.MapFile{Data: []byte(`id: prompt
+version: "2"
+messages:
+  - role: user
+    content: second
+output:
+  format: text
+  validation_mode: none
+`)},
+	})
+	if result, err := ambiguous.InspectPrompt(context.Background(), "prompt", ""); result != nil ||
+		!errors.Is(err, promptkit.ErrPromptLoad) {
+		t.Fatalf("ambiguous prompt result=(%#v, %v), want ErrPromptLoad", result, err)
+	}
+
+	for name, source := range map[string]fstest.MapFS{
+		"malformed definition": {
+			"broken.yaml": &fstest.MapFile{Data: []byte("id: broken\nversion: \"1\"\nunknown: value\n")},
+		},
+		"missing content file": {
+			"broken.yaml": &fstest.MapFile{Data: []byte(`id: broken
+version: "1"
+messages:
+  - role: user
+    content_file: missing.md
+output:
+  format: text
+  validation_mode: none
+`)},
+		},
+	} {
+		t.Run(name, func(t *testing.T) {
+			if result, err := newEngine(t, source).InspectPrompt(context.Background(), "broken", "1"); result != nil ||
+				!errors.Is(err, promptkit.ErrPromptLoad) {
+				t.Fatalf("broken prompt result=(%#v, %v), want ErrPromptLoad", result, err)
+			}
+		})
+	}
+
+	countingFS := &inspectionCountingFS{}
+	canceled, err := promptkit.NewEngine(promptkit.Config{}, promptkit.WithPromptFS(countingFS, "."))
+	if err != nil {
+		t.Fatalf("construct canceled prompt-inspection engine: %v", err)
+	}
+	ctx, cancel := context.WithCancel(context.Background())
+	cancel()
+	if result, err := canceled.InspectPrompt(ctx, "prompt", "1"); result != nil ||
+		!errors.Is(err, promptkit.ErrPromptLoad) || !errors.Is(err, context.Canceled) || countingFS.opens.Load() != 0 {
+		t.Fatalf("canceled inspection result=(%#v, %v), opens=%d", result, err, countingFS.opens.Load())
+	}
+}
+
+func TestInspectPromptReturnsIndependentMetadataMatchingPreparation(t *testing.T) {
+	engine, err := promptkit.NewEngine(promptkit.Config{},
+		promptkit.WithPromptFS(fstest.MapFS{
+			"prompt.yaml": &fstest.MapFile{Data: []byte(`id: prompt
+version: "1"
+default_profile: profile
+inputs:
+  - name: subject
+    content_type: text/plain
+    description: Summary subject.
+messages:
+  - role: user
+    content: summarize
+output:
+  format: markdown
+  validation_mode: basic
+`)},
+		}, "."),
+		promptkit.WithProfiles(promptkit.Profile{
+			ID: "profile", Endpoint: "http://profile.example/v1", Model: "model",
+		}),
+	)
+	if err != nil {
+		t.Fatalf("construct prompt-inspection engine: %v", err)
+	}
+
+	first, err := engine.InspectPrompt(context.Background(), "prompt", "1")
+	if err != nil {
+		t.Fatalf("first inspection: %v", err)
+	}
+	first.Inputs[0].Name = "changed"
+	first.OutputContract.SchemaPath = "changed.json"
+
+	second, err := engine.InspectPrompt(context.Background(), "prompt", "1")
+	if err != nil {
+		t.Fatalf("second inspection: %v", err)
+	}
+	prepared, err := engine.Prepare(context.Background(), promptkit.RunRequest{PromptID: "prompt", PromptVersion: "1"})
+	if err != nil {
+		t.Fatalf("prepare after inspection mutation: %v", err)
+	}
+	if second.Inputs[0].Name != "subject" || second.OutputContract.SchemaPath != "" ||
+		prepared.OutputContract.SchemaPath != "" || second.PromptHash != prepared.PromptHash {
+		t.Fatalf("inspection mutation reached engine-owned prompt metadata: inspection=%#v prepared=%#v", second, prepared)
+	}
+}
+
 func TestPreparedRunJSONOmitsZeroTimingValues(t *testing.T) {
 	payload, err := json.Marshal(promptkit.PreparedRun{})
 	if err != nil {
@@ -138,6 +514,49 @@ func TestUnknownProfileBackendHasProfileLoadIdentity(t *testing.T) {
 	}
 }
 
+func TestLocalBackendConstructsAndRegistersConventionalBackend(t *testing.T) {
+	const (
+		localBackendID = "local"
+		limit          = 2
+	)
+	if promptkit.BackendLocal != localBackendID {
+		t.Fatalf("BackendLocal=%q, want %q", promptkit.BackendLocal, localBackendID)
+	}
+
+	endpoint := "http://local.example/v1"
+	backend := promptkit.LocalBackend(endpoint, limit)
+	want := promptkit.Backend{
+		ID:               localBackendID,
+		Endpoint:         endpoint,
+		ConcurrencyLimit: limit,
+	}
+	if !reflect.DeepEqual(backend, want) {
+		t.Fatalf("LocalBackend()=%+v, want %+v", backend, want)
+	}
+
+	engine, err := promptkit.NewEngine(promptkit.Config{},
+		promptkit.WithPromptFS(contractPromptFS("prompt", "local-profile", "message"), "."),
+		promptkit.WithBackend(backend),
+		promptkit.WithProfiles(promptkit.Profile{
+			ID:        "local-profile",
+			BackendID: localBackendID,
+			Model:     "model",
+		}),
+	)
+	if err != nil {
+		t.Fatalf("construct engine with local backend: %v", err)
+	}
+
+	prepared, err := engine.Prepare(context.Background(), promptkit.RunRequest{PromptID: "prompt"})
+	if err != nil {
+		t.Fatalf("prepare with local backend: %v", err)
+	}
+	if prepared.SelectedBackendID != localBackendID ||
+		prepared.EffectiveModelParams.Endpoint != endpoint {
+		t.Fatalf("unexpected local backend preparation: %+v", prepared)
+	}
+}
+
 func TestCustomBackendFlowsThroughProfilesOverridesAndInjectedClient(t *testing.T) {
 	t.Setenv("CUSTOM_LLM_KEY", "test-key")
 	client := &fakeLLMClient{response: &promptkit.GenerateResponse{Content: "ok"}}
@@ -554,6 +973,24 @@ func TestRepeatedOptionsUseLastValueInEachCategory(t *testing.T) {
 		}
 	})
 
+	t.Run("fallback profile source", func(t *testing.T) {
+		engine, err := promptkit.NewEngine(promptkit.Config{},
+			promptkit.WithPromptFS(contractPromptFS("prompt", "profile", "message"), "."),
+			promptkit.WithFallbackProfileFS(contractProfileFS("profile", "first-model"), "."),
+			promptkit.WithFallbackProfileFS(contractProfileFS("profile", "second-model"), "."),
+		)
+		if err != nil {
+			t.Fatalf("construct engine: %v", err)
+		}
+		prepared, err := engine.Prepare(context.Background(), promptkit.RunRequest{PromptID: "prompt"})
+		if err != nil {
+			t.Fatalf("prepare from last fallback profile source: %v", err)
+		}
+		if prepared.EffectiveModelParams.Model != "second-model" {
+			t.Fatalf("expected last fallback profile source, got %q", prepared.EffectiveModelParams.Model)
+		}
+	})
+
 	t.Run("in-memory profiles", func(t *testing.T) {
 		first := profile
 		first.Model = "first-model"
@@ -638,6 +1075,212 @@ func TestRepeatedOptionsUseLastValueInEachCategory(t *testing.T) {
 	})
 }
 
+func TestFallbackProfileSourcePrecedence(t *testing.T) {
+	const profileID = "application-profile"
+
+	prepareModel := func(t *testing.T, engine *promptkit.Engine, promptID string) string {
+		t.Helper()
+		prepared, err := engine.Prepare(context.Background(), promptkit.RunRequest{PromptID: promptID})
+		if err != nil {
+			t.Fatalf("prepare: %v", err)
+		}
+		return prepared.EffectiveModelParams.Model
+	}
+
+	t.Run("in-memory profiles override ordinary and fallback profiles", func(t *testing.T) {
+		engine, err := promptkit.NewEngine(promptkit.Config{},
+			promptkit.WithPromptFS(contractPromptFS("prompt", profileID, "message"), "."),
+			promptkit.WithFallbackProfileFS(contractProfileFS(profileID, "fallback-model"), "."),
+			promptkit.WithProfileFS(contractProfileFS(profileID, "ordinary-model"), "."),
+			promptkit.WithProfiles(promptkit.Profile{ID: profileID, Endpoint: "http://example.test/v1", Model: "memory-model"}),
+		)
+		if err != nil {
+			t.Fatalf("construct engine: %v", err)
+		}
+		if model := prepareModel(t, engine, "prompt"); model != "memory-model" {
+			t.Fatalf("expected in-memory profile, got %q", model)
+		}
+	})
+
+	t.Run("ordinary filesystem source overrides fallback profile", func(t *testing.T) {
+		engine, err := promptkit.NewEngine(promptkit.Config{},
+			promptkit.WithPromptFS(contractPromptFS("prompt", profileID, "message"), "."),
+			promptkit.WithFallbackProfileFS(contractProfileFS(profileID, "fallback-model"), "."),
+			promptkit.WithProfileFS(contractProfileFS(profileID, "ordinary-model"), "."),
+		)
+		if err != nil {
+			t.Fatalf("construct engine: %v", err)
+		}
+		if model := prepareModel(t, engine, "prompt"); model != "ordinary-model" {
+			t.Fatalf("expected ordinary profile, got %q", model)
+		}
+	})
+
+	t.Run("configured directory overrides fallback profile", func(t *testing.T) {
+		profileDir := t.TempDir()
+		writePublicProfileFile(t, profileDir, profileID, "http://example.test/v1", "directory-model")
+		engine, err := promptkit.NewEngine(promptkit.Config{ProfileDir: profileDir},
+			promptkit.WithPromptFS(contractPromptFS("prompt", profileID, "message"), "."),
+			promptkit.WithFallbackProfileFS(contractProfileFS(profileID, "fallback-model"), "."),
+		)
+		if err != nil {
+			t.Fatalf("construct engine: %v", err)
+		}
+		if model := prepareModel(t, engine, "prompt"); model != "directory-model" {
+			t.Fatalf("expected configured directory profile, got %q", model)
+		}
+	})
+
+	t.Run("fallback profile overrides built-in profile", func(t *testing.T) {
+		engine, err := promptkit.NewEngine(promptkit.Config{},
+			promptkit.WithPromptFS(contractPromptFS("prompt", "mistral-small-3", "message"), "."),
+			promptkit.WithFallbackProfileFS(contractProfileFS("mistral-small-3", "fallback-model"), "."),
+		)
+		if err != nil {
+			t.Fatalf("construct engine: %v", err)
+		}
+		if model := prepareModel(t, engine, "prompt"); model != "fallback-model" {
+			t.Fatalf("expected fallback profile, got %q", model)
+		}
+	})
+
+	t.Run("missing fallback profile uses built-in profile", func(t *testing.T) {
+		t.Setenv("OPENROUTER_API_KEY", "test-key")
+		baseline, err := promptkit.NewEngine(promptkit.Config{},
+			promptkit.WithPromptFS(contractPromptFS("prompt", "mistral-small-3", "message"), "."),
+		)
+		if err != nil {
+			t.Fatalf("construct baseline engine: %v", err)
+		}
+		want := prepareModel(t, baseline, "prompt")
+
+		engine, err := promptkit.NewEngine(promptkit.Config{},
+			promptkit.WithPromptFS(contractPromptFS("prompt", "mistral-small-3", "message"), "."),
+			promptkit.WithFallbackProfileFS(contractProfileFS(profileID, "fallback-model"), "."),
+		)
+		if err != nil {
+			t.Fatalf("construct engine: %v", err)
+		}
+		if model := prepareModel(t, engine, "prompt"); model != want {
+			t.Fatalf("expected built-in profile model %q, got %q", want, model)
+		}
+	})
+}
+
+func TestFallbackProfileSourcePreservesLazyLoadingAndErrors(t *testing.T) {
+	const profileID = "application-profile"
+
+	t.Run("construction defers malformed fallback profiles", func(t *testing.T) {
+		_, err := promptkit.NewEngine(promptkit.Config{},
+			promptkit.WithPromptFS(fstest.MapFS{}, "."),
+			promptkit.WithFallbackProfileFS(fstest.MapFS{
+				"broken.yaml": &fstest.MapFile{Data: []byte("id: broken\nunknown: value\n")},
+			}, "."),
+		)
+		if err != nil {
+			t.Fatalf("construct engine with malformed fallback profile: %v", err)
+		}
+	})
+
+	t.Run("unrelated malformed fallback profile does not block matching definition", func(t *testing.T) {
+		engine, err := promptkit.NewEngine(promptkit.Config{},
+			promptkit.WithPromptFS(contractPromptFS("prompt", profileID, "message"), "."),
+			promptkit.WithFallbackProfileFS(fstest.MapFS{
+				"broken.yaml": &fstest.MapFile{Data: []byte("id: unrelated\nunknown: value\n")},
+				"valid.yaml":  &fstest.MapFile{Data: []byte("id: application-profile\nendpoint: http://example.test/v1\nmodel: fallback-model\n")},
+			}, "."),
+		)
+		if err != nil {
+			t.Fatalf("construct engine: %v", err)
+		}
+		prepared, err := engine.Prepare(context.Background(), promptkit.RunRequest{PromptID: "prompt"})
+		if err != nil {
+			t.Fatalf("prepare from valid fallback profile: %v", err)
+		}
+		if prepared.EffectiveModelParams.Model != "fallback-model" {
+			t.Fatalf("unexpected fallback profile model: %q", prepared.EffectiveModelParams.Model)
+		}
+	})
+
+	t.Run("matching malformed fallback profile does not reach built-in profile", func(t *testing.T) {
+		engine, err := promptkit.NewEngine(promptkit.Config{},
+			promptkit.WithPromptFS(contractPromptFS("prompt", "mistral-small-3", "message"), "."),
+			promptkit.WithFallbackProfileFS(fstest.MapFS{
+				"mistral-small-3.yaml": &fstest.MapFile{Data: []byte("id: mistral-small-3\nendpoint: http://example.test/v1\nmodel: fallback-model\nunknown: value\n")},
+			}, "."),
+		)
+		if err != nil {
+			t.Fatalf("construct engine: %v", err)
+		}
+		if _, err := engine.Prepare(context.Background(), promptkit.RunRequest{PromptID: "prompt"}); !errors.Is(err, promptkit.ErrProfileLoad) {
+			t.Fatalf("expected ErrProfileLoad, got %v", err)
+		}
+	})
+
+	t.Run("matching malformed ordinary profile does not reach fallback profile", func(t *testing.T) {
+		engine, err := promptkit.NewEngine(promptkit.Config{},
+			promptkit.WithPromptFS(contractPromptFS("prompt", profileID, "message"), "."),
+			promptkit.WithProfileFS(fstest.MapFS{
+				"application-profile.yaml": &fstest.MapFile{Data: []byte("id: application-profile\nendpoint: http://example.test/v1\nmodel: ordinary-model\nunknown: value\n")},
+			}, "."),
+			promptkit.WithFallbackProfileFS(contractProfileFS(profileID, "fallback-model"), "."),
+		)
+		if err != nil {
+			t.Fatalf("construct engine: %v", err)
+		}
+		if _, err := engine.Prepare(context.Background(), promptkit.RunRequest{PromptID: "prompt"}); !errors.Is(err, promptkit.ErrProfileLoad) {
+			t.Fatalf("expected ErrProfileLoad, got %v", err)
+		}
+	})
+}
+
+func TestFallbackProfileSourceWorksAcrossWorkflows(t *testing.T) {
+	const profileID = "application-profile"
+	client := &fakeLLMClient{response: &promptkit.GenerateResponse{Content: "ok"}}
+	engine, err := promptkit.NewEngine(promptkit.Config{},
+		promptkit.WithPromptFS(contractPromptFS("prompt", profileID, "message"), "."),
+		promptkit.WithFallbackProfileFS(contractProfileFS(profileID, "fallback-model"), "."),
+		promptkit.WithLLMClient(client),
+	)
+	if err != nil {
+		t.Fatalf("construct engine: %v", err)
+	}
+
+	inspection, err := engine.InspectProfile(context.Background(), profileID)
+	if err != nil {
+		t.Fatalf("inspect fallback profile: %v", err)
+	}
+	prepared, err := engine.Prepare(context.Background(), promptkit.RunRequest{PromptID: "prompt"})
+	if err != nil {
+		t.Fatalf("prepare fallback profile: %v", err)
+	}
+	preparedExecution, err := engine.PrepareExecution(context.Background(), promptkit.RunRequest{PromptID: "prompt"})
+	if err != nil {
+		t.Fatalf("prepare execution with fallback profile: %v", err)
+	}
+	preparedDetails := preparedExecution.Details()
+	preparedResult, err := engine.RunPrepared(context.Background(), preparedExecution)
+	if err != nil {
+		t.Fatalf("run prepared fallback profile: %v", err)
+	}
+	runResult, err := engine.Run(context.Background(), promptkit.RunRequest{PromptID: "prompt"})
+	if err != nil {
+		t.Fatalf("run fallback profile: %v", err)
+	}
+
+	for name, model := range map[string]string{
+		"inspection":         inspection.EffectiveModelParams.Model,
+		"preparation":        prepared.EffectiveModelParams.Model,
+		"prepared execution": preparedDetails.EffectiveModelParams.Model,
+		"prepared result":    preparedResult.EffectiveModelParams.Model,
+		"run result":         runResult.EffectiveModelParams.Model,
+	} {
+		if model != "fallback-model" {
+			t.Fatalf("%s model=%q, want fallback-model", name, model)
+		}
+	}
+}
+
 func TestEngineSupportsConcurrentPrepareAndRun(t *testing.T) {
 	engine, err := promptkit.NewEngine(promptkit.Config{},
 		promptkit.WithPromptFS(contractPromptFS("prompt", "profile", "message"), "."),

types.go

@@ -51,7 +51,8 @@ const (
 	// ValidationPassed means the generated output satisfied its contract.
 	ValidationPassed ValidationStatus = "passed"
 	// ValidationFailed means validation completed and rejected the generated
-	// output. Engine.Run returns this status in a result, not as an error.
+	// output. Engine.Run and Engine.RunPrepared return this status in a result,
+	// not as an error.
 	ValidationFailed ValidationStatus = "failed"
 	// ValidationSkipped means ValidationNone selected no content check.
 	ValidationSkipped ValidationStatus = "skipped"
@@ -78,9 +79,10 @@ const (
 // RunRequest selects one prompt execution. It has no stable JSON
 // representation.
 //
-// Prepare and Run copy the request's maps, pointers, and nested
-// JSON-compatible values before using them. The caller may mutate the request
-// after either method returns.
+// Prepare, PrepareExecution, and Run copy the request's maps, pointers, and
+// nested JSON-compatible values before using them. The caller may mutate the
+// request after any method returns. A successful PrepareExecution retains its
+// own private execution snapshot for RunPrepared.
 type RunRequest struct {
 	// PromptID is the required non-empty prompt identifier.
 	PromptID string
@@ -98,12 +100,14 @@ type RunRequest struct {
 	// opaque consumer metadata, not a credential, and may be exposed in
 	// prepared values, results, collaborator requests, provider requests, and
 	// provider observability. Callers should use stable, non-sensitive
-	// identifiers. An overlong direct value makes Prepare or Run return an
-	// error matching ErrInvalidRequest.
+	// identifiers. An overlong direct value makes Prepare, PrepareExecution, or
+	// Run return an error matching ErrInvalidRequest.
 	SessionID string
 	// APIKey is a request-scoped direct credential. It takes precedence over
 	// APIKeyEnv, is passed to the selected LLMClient, and is never included in
-	// prepared values, results, hashes, JSON, String, or GoString output.
+	// prepared values, results, hashes, JSON, String, or GoString output. A
+	// successful PrepareExecution retains it only in the opaque handle until
+	// RunPrepared claims the handle or Discard invalidates it.
 	APIKey string `json:"-"`
 	// Inputs maps prompt input names to references. A nil or empty map is valid
 	// only when the selected prompt and its templates require no inputs.
@@ -112,16 +116,18 @@ type RunRequest struct {
 	// empty maps are equivalent.
 	Vars map[string]string
 	// Execution optionally overrides individual execution settings. Nil uses
-	// the selected profile over its backend, when any, and framework defaults.
+	// the selected profile over its backend, when any, and the framework
+	// baseline.
 	Execution *ExecutionTargetOverride
 	// Validation optionally replaces the prompt's complete output contract. It
 	// does not merge individual fields. Nil uses the prompt contract.
 	Validation *OutputContract
 }
 
-// PreparedRun contains prepared prompt execution state. It does not include
-// resolved API key values, model output, validation results, or internal target
-// presence metadata. PreparedRun has a stable JSON representation.
+// PreparedRun contains prepared prompt execution state returned by
+// [Engine.Prepare] or [PreparedExecution.Details]. It does not include resolved
+// API key values, model output, validation results, or internal target presence
+// metadata. PreparedRun has a stable JSON representation.
 //
 // All maps, slices, pointers, and schema values are caller-owned copies. JSON
 // timestamps use RFC 3339 and zero timing values are omitted. Hash formats are
@@ -139,9 +145,10 @@ type PreparedRun struct {
 	// SelectedBackendID equals EffectiveModelParams.BackendID. It is empty for
 	// an endpoint-only profile.
 	SelectedBackendID string `json:"selected_backend_id,omitempty"`
-	// EffectiveModelParams contains framework defaults overlaid by the selected
-	// backend, profile, and then request overrides. It excludes resolved API-key
-	// values.
+	// EffectiveModelParams contains settings resolved from the framework timeout
+	// baseline, selected backend, profile, and then request overrides. Unset
+	// optional provider controls remain zero rather than reporting a provider
+	// default. It excludes resolved API-key values.
 	EffectiveModelParams ExecutionTarget `json:"effective_model_params"`
 	// OutputContract is the complete effective output contract.
 	OutputContract OutputContract `json:"output_contract"`
@@ -154,7 +161,8 @@ type PreparedRun struct {
 	SessionID string `json:"session_id,omitempty"`
 	// RenderedPromptHash is an opaque equality value for SessionID and Messages.
 	RenderedPromptHash string `json:"rendered_prompt_hash"`
-	// Messages are the rendered messages that Run passes to the LLM client.
+	// Messages are the rendered messages that Run or RunPrepared passes to the
+	// LLM client.
 	Messages []RenderedMessage `json:"messages"`
 	// StartTime is the UTC time at which preparation began.
 	StartTime time.Time `json:"start_time,omitempty"`
@@ -214,12 +222,15 @@ type RunResult struct {
 	InputHashes map[string]string `json:"input_hashes,omitempty"`
 	// Usage is the token accounting reported by the LLM client.
 	Usage TokenUsage `json:"usage"`
-	// StartTime is the UTC time immediately before preparation begins.
+	// StartTime is the UTC time immediately before ordinary Run preparation or
+	// after RunPrepared claims its handle.
 	StartTime time.Time `json:"start_time,omitempty"`
 	// EndTime is the UTC time after generation and validation complete.
 	EndTime time.Time `json:"end_time,omitempty"`
-	// Duration covers preparation, generation, and validation. JSON represents
-	// it as integer milliseconds in duration_ms and omits a zero value.
+	// Duration covers preparation, generation, and validation for Run. For
+	// RunPrepared it covers only the execution attempt after claim and excludes
+	// preparation and consumer-held delay. JSON represents it as integer
+	// milliseconds in duration_ms and omits a zero value.
 	Duration time.Duration `json:"-"`
 }
 
@@ -259,10 +270,11 @@ type Artifact struct {
 // ArtifactReader resolves a prompt input reference into its content.
 //
 // Read may be called concurrently. It must honor ctx cancellation to make
-// Prepare and Run responsive to cancellation. The engine passes a copied ref
-// and immediately copies the returned Artifact.Body; it does not retain either
-// value. Readers supply artifact metadata, and the engine assigns an input-map
-// name only when the returned artifact name is empty.
+// Prepare, PrepareExecution, and Run responsive to cancellation. The engine
+// passes a copied ref and immediately copies the returned Artifact.Body; it
+// does not retain either value. Readers supply artifact metadata, and the
+// engine assigns an input-map name only when the returned artifact name is
+// empty.
 //
 // An injected reader owns any application-specific path containment,
 // authorization, content-size, and content-type policy. It must protect
@@ -288,13 +300,17 @@ type ExecutionTarget struct {
 	Endpoint string `json:"endpoint"`
 	// Model is the provider model identifier.
 	Model string `json:"model"`
-	// Temperature is the effective sampling temperature from 0 through 2.
+	// Temperature is the resolved sampling temperature from 0 through 2. Zero
+	// leaves the field unspecified to compatible providers unless the
+	// corresponding ExecutionTargetPresence bit is true.
 	Temperature float64 `json:"temperature"`
-	// MaxTokens is the non-negative effective output-token limit. Zero leaves
-	// the limit unspecified to compatible providers unless it was an explicit
-	// request override.
+	// MaxTokens is the non-negative resolved output-token limit. Zero leaves
+	// the limit unspecified to compatible providers unless the corresponding
+	// ExecutionTargetPresence bit is true.
 	MaxTokens int `json:"max_tokens"`
-	// TopP is the effective nucleus-sampling value from 0 through 1.
+	// TopP is the resolved nucleus-sampling value from 0 through 1. Zero leaves
+	// the field unspecified to compatible providers unless the corresponding
+	// ExecutionTargetPresence bit is true.
 	TopP float64 `json:"top_p"`
 	// TimeoutSeconds is the non-negative per-generation deadline. Zero disables
 	// this deadline without disabling caller cancellation or the transport cap.
@@ -310,6 +326,63 @@ type ExecutionTarget struct {
 	ExtraParams map[string]any `json:"extra_params"`
 }
 
+// ProfileInspection is the caller-owned result of [Engine.InspectProfile].
+// It has no stable JSON representation.
+//
+// EffectiveModelParams contains a copied effective target. APIKeyRequired is
+// separate from that target to preserve ExecutionTarget's general execution
+// and stable JSON contracts.
+type ProfileInspection struct {
+	// ProfileID is the trimmed, exact profile ID inspected by the engine.
+	ProfileID string
+	// EffectiveModelParams contains settings resolved from the framework timeout
+	// baseline, selected backend, and then profile, without a request override.
+	// Unset optional provider controls remain zero rather than reporting a
+	// provider default. APIKeyEnv is an environment-variable name, never its
+	// credential value.
+	EffectiveModelParams ExecutionTarget
+	// APIKeyRequired reports that a later execution must supply a direct API
+	// key or an explicit request environment override. It is mutually exclusive
+	// with a nonblank EffectiveModelParams.APIKeyEnv.
+	APIKeyRequired bool
+}
+
+// PromptInputDefinition describes one declared prompt input.
+// It has no stable JSON representation.
+type PromptInputDefinition struct {
+	// Name is the normalized prompt input name.
+	Name string
+	// Required is the prompt definition's declared required flag. When true,
+	// preparation fails if the input is omitted. A false value does not account
+	// for input references in message or session-ID templates.
+	Required bool
+	// ContentType is the declared input media-type metadata.
+	ContentType string
+	// Description is the declared human-readable input description.
+	Description string
+}
+
+// PromptInspection is the caller-owned result of [Engine.InspectPrompt].
+// It has no stable JSON representation.
+//
+// Inputs contains copied declared input metadata in definition order.
+// OutputContract is the normalized contract declared by the prompt definition,
+// rather than a request-level effective override. PromptHash is opaque.
+type PromptInspection struct {
+	// PromptID is the normalized ID of the selected prompt definition.
+	PromptID string
+	// PromptVersion is the normalized version of the selected prompt definition.
+	PromptVersion string
+	// PromptHash is the opaque equality value for the selected definition.
+	PromptHash string
+	// DefaultProfileID is declared metadata and is not resolved by inspection.
+	DefaultProfileID string
+	// Inputs contains caller-owned declared input metadata in definition order.
+	Inputs []PromptInputDefinition
+	// OutputContract is the normalized contract declared by the definition.
+	OutputContract OutputContract
+}
+
 // ExecutionTargetOverride represents per-request runtime setting overrides and
 // has no stable JSON representation.
 //
@@ -317,19 +390,26 @@ type ExecutionTarget struct {
 // fields replace profile values and preserve explicit zero or empty values. A
 // non-empty ExtraParams map replaces the complete profile or backend map
 // rather than merging keys. Empty string fields, nil pointers, and a nil or
-// empty ExtraParams map inherit the selected profile over its backend, when
-// any, and framework defaults.
+// empty ExtraParams map inherit lower-precedence values. An optional provider
+// control that remains zero is unspecified; TimeoutSeconds retains its
+// framework deadline when no higher-precedence value is present.
 type ExecutionTargetOverride struct {
 	// Endpoint replaces the profile or backend endpoint when non-empty without
 	// changing the effective BackendID.
 	Endpoint string
 	// Model replaces the profile model when non-empty.
 	Model string
-	// Temperature, when non-nil, must point to a value from 0 through 2.
+	// Temperature, when non-nil, must point to a value from 0 through 2. A
+	// pointed-to zero is explicitly present; nil inherits a lower-precedence
+	// value and otherwise leaves the provider control unspecified.
 	Temperature *float64
-	// MaxTokens, when non-nil, must point to a non-negative value.
+	// MaxTokens, when non-nil, must point to a non-negative value. A pointed-to
+	// zero is explicitly present; nil inherits a lower-precedence value and
+	// otherwise leaves the provider control unspecified.
 	MaxTokens *int
-	// TopP, when non-nil, must point to a value from 0 through 1.
+	// TopP, when non-nil, must point to a value from 0 through 1. A pointed-to
+	// zero is explicitly present; nil inherits a lower-precedence value and
+	// otherwise leaves the provider control unspecified.
 	TopP *float64
 	// TimeoutSeconds, when non-nil, must point to a non-negative value. A
 	// pointed-to zero disables the per-generation deadline.
@@ -360,9 +440,12 @@ type ExecutionTargetOverride struct {
 // use profile YAML api_key_env with file and FS profile sources. Profile has no
 // stable JSON representation.
 //
-// WithProfiles validates and copies Profile values during NewEngine. Numeric
-// zero, blank strings, and an empty ExtraParams map inherit framework defaults;
-// use ExecutionTargetOverride pointer fields to request explicit numeric zero.
+// WithProfiles validates and copies Profile values during NewEngine. Zero
+// Temperature, MaxTokens, and TopP values and blank ServiceTier and
+// ReasoningEffort values leave those provider controls unspecified. A zero
+// TimeoutSeconds retains the framework deadline, while an empty ExtraParams map
+// inherits backend request defaults. Use ExecutionTargetOverride pointer fields
+// to request an explicit numeric zero.
 type Profile struct {
 	// ID is the required non-blank profile identifier. WithProfiles trims it.
 	ID string
@@ -376,19 +459,19 @@ type Profile struct {
 	Endpoint string
 	// Model is the required non-blank provider model identifier.
 	Model string
-	// Temperature is from 0 through 2. Zero inherits the framework default.
+	// Temperature is from 0 through 2. Zero leaves the provider control
+	// unspecified.
 	Temperature float64
-	// MaxTokens is non-negative. Zero inherits the framework default.
+	// MaxTokens is non-negative. Zero leaves the provider control unspecified.
 	MaxTokens int
-	// TopP is from 0 through 1. Zero inherits the framework default rather than
-	// selecting an explicit zero.
+	// TopP is from 0 through 1. Zero leaves the provider control unspecified
+	// rather than selecting an explicit zero.
 	TopP float64
-	// TimeoutSeconds is non-negative. Zero inherits the framework default.
+	// TimeoutSeconds is non-negative. Zero retains the framework deadline.
 	TimeoutSeconds int
-	// ServiceTier is optional; a blank value inherits the framework default.
+	// ServiceTier is optional; a blank value leaves it unspecified.
 	ServiceTier string
-	// ReasoningEffort is optional; a blank value inherits the framework
-	// default.
+	// ReasoningEffort is optional; a blank value leaves it unspecified.
 	ReasoningEffort string
 	// APIKeyRequired clears a backend's inherited API-key environment name and
 	// requires a non-blank RunRequest.APIKey unless the request explicitly
@@ -559,25 +642,26 @@ type StructuredOutputJSONSpec struct {
 	Schema any `json:"schema"`
 }
 
-// LLMClient executes rendered prompts for [Engine.Run].
+// LLMClient executes rendered prompts for [Engine.Run] and
+// [Engine.RunPrepared].
 //
 // Generate is scheduled according to the resolved backend's capacity policy.
 // It may still be called concurrently for different backend pools or unlimited
 // backends. Cancellation while waiting for capacity can prevent Generate from
 // being called. Once invoked, it must honor context cancellation to make Run
-// responsive to cancellation. The request and all nested maps, slices, and
-// pointers are client-owned copies and may be mutated or retained without
-// affecting engine state.
+// and RunPrepared responsive to cancellation. The request and all nested maps,
+// slices, and pointers are client-owned copies and may be mutated or retained
+// without affecting engine state.
 //
 // Generate receives rendered messages and may receive a direct API key. A
 // client must protect those values and any raw output in its logging, storage,
 // and retained copies. It is responsible for the cancellation behavior of any
 // work it starts and for synchronizing access to retained or shared data.
 //
-// A returned error makes Run return ErrLLMGenerate while preserving the client
-// error through errors.Is. A nil response with a nil error also produces
-// ErrLLMGenerate. Promptkit copies the non-nil response before returning from
-// Run.
+// A returned error makes Run or RunPrepared return ErrLLMGenerate while
+// preserving the client error through errors.Is. A nil response with a nil
+// error also produces ErrLLMGenerate. Promptkit copies the non-nil response
+// before returning from either method.
 type LLMClient interface {
 	Generate(context.Context, GenerateRequest) (*GenerateResponse, error)
 }