diff --git a/docs/consumers/pkg-promptkit.md b/docs/consumers/pkg-promptkit.md
index e723395..5b57d96 100644
--- a/docs/consumers/pkg-promptkit.md
+++ b/docs/consumers/pkg-promptkit.md
@@ -558,6 +558,14 @@ disclosure policy. Promptkit does not assign retry or presentation behavior.
 The [`GenerationError` GoDoc](../../generation_error.go) owns the exact typed
 error contract.
 
+Generation failures also match `ErrTransport` when the built-in HTTP request
+fails, or `ErrMalformedResponse` when a successful provider response cannot be
+decoded or an injected client returns a nil successful response. Both retain
+`ErrLLMGenerate`; transport failures retain their original causes for
+`errors.Is` and `errors.As`. Caller cancellation and deadlines retain their
+standard context identities. Arbitrary injected-client errors retain their own
+identities without Promptkit guessing a failure category from their text.
+
 ## Application Boundary
 
 Promptkit is an importable library. It does not own a command, inbound HTTP
diff --git a/docs/roadmap/implementation.md b/docs/roadmap/implementation.md
index 61f095b..434e177 100644
--- a/docs/roadmap/implementation.md
+++ b/docs/roadmap/implementation.md
@@ -216,6 +216,8 @@ admission slot. No asynchronous delivery or durable history.
 
 ## Stage 1 — Expose Missing Generation Failure Categories
 
+**Completed:** 2026-09-23.
+
 Read `errors.go`, `generation_error.go`, `llm_adapter.go`, the LLM error types,
 and public error-boundary tests. Add public transport/malformed sentinels and
 map current built-in errors while preserving cause identity and redacted text.
diff --git a/engine.go b/engine.go
index 7b470b4..04a28dd 100644
--- a/engine.go
+++ b/engine.go
@@ -61,6 +61,15 @@ var (
 	// [GenerationError]. Errors returned by an injected LLMClient remain
 	// available through errors.Is.
 	ErrLLMGenerate = errors.New("failed to generate output")
+	// ErrTransport identifies a failed HTTP request by the built-in model
+	// client. It also matches ErrLLMGenerate and retains the transport cause
+	// for errors.Is and errors.As. It does not imply that the provider received
+	// no work.
+	ErrTransport = errors.New("model transport failed")
+	// ErrMalformedResponse identifies an invalid successful model response or
+	// a nil successful response from an injected client. It also matches
+	// ErrLLMGenerate.
+	ErrMalformedResponse = errors.New("malformed model response")
 	// ErrValidation identifies an operational failure to load or compile a
 	// schema or validate output. A completed validation whose Status is
 	// ValidationFailed is returned in RunResult without this error.
@@ -483,7 +492,10 @@ func (e *Engine) PrepareExecution(ctx context.Context, req RunRequest) (*Prepare
 // no partial result.
 //
 // Run can return every error category documented by [Engine.Prepare], plus
-// ErrCapacityExceeded and ErrLLMGenerate. An engine admission rejection is
+// ErrCapacityExceeded and ErrLLMGenerate. Built-in HTTP execution failures
+// additionally match ErrTransport; malformed successful responses and nil
+// successful injected responses additionally match ErrMalformedResponse.
+// 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
@@ -532,7 +544,9 @@ func (e *Engine) Run(ctx context.Context, req RunRequest) (*RunResult, error) {
 // preserving documented collaborator and context identities. An engine
 // admission rejection is discoverable as [CapacityError] and still matches
 // ErrCapacityExceeded. A built-in OpenAI-compatible non-2xx response is
-// discoverable as [GenerationError]. A completed content-validation rejection,
+// discoverable as [GenerationError]. Built-in HTTP execution failures match
+// ErrTransport; malformed successful responses and nil successful injected
+// responses match ErrMalformedResponse. A completed content-validation rejection,
 // including repair exhaustion, 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) {
diff --git a/errors.go b/errors.go
index 612d0dc..014bd31 100644
--- a/errors.go
+++ b/errors.go
@@ -38,7 +38,14 @@ func mapPublicError(err error) error {
 	if publicErr == nil {
 		return err
 	}
-	return fmt.Errorf("%w: %w", publicErr, err)
+	switch {
+	case errors.Is(err, llm.ErrTransport):
+		return fmt.Errorf("%w: %w: %w", publicErr, ErrTransport, err)
+	case errors.Is(err, llm.ErrMalformedResponse):
+		return fmt.Errorf("%w: %w: %w", publicErr, ErrMalformedResponse, err)
+	default:
+		return fmt.Errorf("%w: %w", publicErr, err)
+	}
 }
 
 func publicErrorFor(err error) error {
diff --git a/errors_internal_test.go b/errors_internal_test.go
index 82d066d..6657949 100644
--- a/errors_internal_test.go
+++ b/errors_internal_test.go
@@ -73,3 +73,10 @@ func TestMapPublicErrorPreservesValidationAroundGenerationError(t *testing.T) {
 		t.Fatalf("mapped error exposes internal ProviderHTTPError: %v", err)
 	}
 }
+
+func TestMapPublicErrorDoesNotTreatRequestEncodingAsTransport(t *testing.T) {
+	err := mapPublicError(fmt.Errorf("%w: %w", usecase.ErrLLMGenerate, llm.ErrRequestFailed))
+	if !errors.Is(err, ErrLLMGenerate) || errors.Is(err, ErrTransport) {
+		t.Fatalf("request encoding category = %v", err)
+	}
+}
diff --git a/generation_error_contract_test.go b/generation_error_contract_test.go
index d1252a7..20ace5f 100644
--- a/generation_error_contract_test.go
+++ b/generation_error_contract_test.go
@@ -94,6 +94,89 @@ func TestBuiltInRepairGenerationError(t *testing.T) {
 	assertGenerationError(t, err, http.StatusUnprocessableEntity, codeMarker, typeMarker, messageMarker)
 }
 
+func TestBuiltInGenerationFailureCategories(t *testing.T) {
+	transportCause := errors.New("private transport diagnostic")
+	config := contractConfig(frameworkSchemaDir)
+	config.HTTPClient = &http.Client{Transport: roundTripFunc(func(*http.Request) (*http.Response, error) {
+		return nil, transportCause
+	})}
+	engine, err := promptkit.NewEngine(config)
+	if err != nil {
+		t.Fatal(err)
+	}
+	result, err := engine.Run(context.Background(), generationErrorRunRequest())
+	if result != nil || !errors.Is(err, promptkit.ErrLLMGenerate) || !errors.Is(err, promptkit.ErrTransport) || !errors.Is(err, transportCause) {
+		t.Fatalf("transport result=%#v, error=%v", result, err)
+	}
+	if errors.Is(err, promptkit.ErrMalformedResponse) || strings.Contains(fmt.Sprintf("%#v", err), transportCause.Error()) {
+		t.Fatalf("transport category or formatting = %v", err)
+	}
+
+	for _, tc := range []struct {
+		name string
+		body string
+	}{
+		{name: "invalid framing", body: `{"choices":[]}{}`},
+		{name: "missing choices", body: `{"choices":[]}`},
+		{name: "missing content", body: `{"choices":[{"message":{}}]}`},
+	} {
+		t.Run(tc.name, func(t *testing.T) {
+			engine := newBuiltInGenerationErrorEngine(t, http.StatusOK, tc.body)
+			result, err := engine.Run(context.Background(), generationErrorRunRequest())
+			if result != nil || !errors.Is(err, promptkit.ErrLLMGenerate) || !errors.Is(err, promptkit.ErrMalformedResponse) || errors.Is(err, promptkit.ErrTransport) {
+				t.Fatalf("malformed result=%#v, error=%v", result, err)
+			}
+		})
+	}
+}
+
+func TestBuiltInGenerationPreservesContextIdentity(t *testing.T) {
+	for _, tc := range []struct {
+		name  string
+		cause error
+	}{
+		{name: "canceled", cause: context.Canceled},
+		{name: "deadline", cause: context.DeadlineExceeded},
+	} {
+		t.Run(tc.name, func(t *testing.T) {
+			config := contractConfig(frameworkSchemaDir)
+			config.HTTPClient = &http.Client{Transport: roundTripFunc(func(*http.Request) (*http.Response, error) {
+				return nil, tc.cause
+			})}
+			engine, err := promptkit.NewEngine(config)
+			if err != nil {
+				t.Fatal(err)
+			}
+			result, err := engine.Run(context.Background(), generationErrorRunRequest())
+			if result != nil || !errors.Is(err, tc.cause) || !errors.Is(err, promptkit.ErrLLMGenerate) || !errors.Is(err, promptkit.ErrTransport) {
+				t.Fatalf("context result=%#v, error=%v", result, err)
+			}
+		})
+	}
+}
+
+func TestInjectedGenerationFailureCategories(t *testing.T) {
+	for _, tc := range []struct {
+		name      string
+		client    *fakeLLMClient
+		malformed bool
+	}{
+		{name: "nil response", client: &fakeLLMClient{}, malformed: true},
+		{name: "arbitrary error", client: &fakeLLMClient{err: errors.New("injected failure")}},
+	} {
+		t.Run(tc.name, func(t *testing.T) {
+			engine := newContractEngineWithOptions(t, frameworkSchemaDir, promptkit.WithLLMClient(tc.client))
+			result, err := engine.Run(context.Background(), generationErrorRunRequest())
+			if result != nil || !errors.Is(err, promptkit.ErrLLMGenerate) || errors.Is(err, promptkit.ErrTransport) || errors.Is(err, promptkit.ErrMalformedResponse) != tc.malformed {
+				t.Fatalf("injected result=%#v, error=%v", result, err)
+			}
+			if tc.client.err != nil && !errors.Is(err, tc.client.err) {
+				t.Fatalf("injected cause lost: %v", err)
+			}
+		})
+	}
+}
+
 func assertGenerationError(t *testing.T, err error, statusCode int, code, providerType, message string) {
 	t.Helper()
 
diff --git a/internal/llm/openai_compatible_client.go b/internal/llm/openai_compatible_client.go
index ba0c760..1892e48 100644
--- a/internal/llm/openai_compatible_client.go
+++ b/internal/llm/openai_compatible_client.go
@@ -21,6 +21,7 @@ var (
 	ErrInvalidConfig     = errors.New("invalid llm client configuration")
 	ErrInvalidRequest    = errors.New("invalid generate request")
 	ErrRequestFailed     = errors.New("llm request failed")
+	ErrTransport         = errors.New("llm transport failed")
 	ErrUnexpectedStatus  = errors.New("llm returned non-success status")
 	ErrMalformedResponse = errors.New("malformed llm response")
 )
@@ -36,7 +37,7 @@ func (e *requestFailedError) Error() string {
 }
 
 func (e *requestFailedError) Unwrap() []error {
-	return []error{ErrRequestFailed, e.cause}
+	return []error{ErrRequestFailed, ErrTransport, e.cause}
 }
 
 type OpenAICompatibleConfig struct {
diff --git a/llm_adapter.go b/llm_adapter.go
index 9efbbcb..d4a2c46 100644
--- a/llm_adapter.go
+++ b/llm_adapter.go
@@ -17,7 +17,7 @@ func (a publicLLMClientAdapter) Generate(ctx context.Context, req domain.Generat
 		return nil, err
 	}
 	if resp == nil {
-		return nil, fmt.Errorf("%w: llm client returned nil response", ErrLLMGenerate)
+		return nil, fmt.Errorf("%w: %w: llm client returned nil response", ErrLLMGenerate, ErrMalformedResponse)
 	}
 	return toDomainGenerateResponse(resp), nil
 }