Expose transport and malformed generation errors

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) {