Bound external result file reads

internal/stage/analyze.go

@@ -916,14 +916,11 @@ func deriveSessionVarValue(name string, session *config.SessionConfig) (string,
 }
 
 func requireNonEmptyFile(path string, label string) error {
-	if err := requireFile(path, label); err != nil {
+	data, err := readExternalResult(path, label)
+	if err != nil {
 		return err
 	}
-	info, err := os.Stat(path)
-	if err != nil {
-		return fmt.Errorf("%s %q stat failed: %w", label, path, err)
-	}
-	if info.Size() <= 0 {
+	if len(data) == 0 {
 		return fmt.Errorf("%s %q is empty", label, path)
 	}
 	return nil
@@ -968,9 +965,9 @@ func resolveRenderDebugEnabled(global bool, perArtifact *bool) bool {
 }
 
 func validateJSONFile(path string) error {
-	data, err := os.ReadFile(path)
+	data, err := readExternalResult(path, "scriptorium artifact result")
 	if err != nil {
-		return fmt.Errorf("read file: %w", err)
+		return err
 	}
 	var payload any
 	if err := json.Unmarshal(data, &payload); err != nil {

internal/stage/external_result.go

@@ -0,0 +1,19 @@
+package stage
+
+import (
+	"fmt"
+
+	"gitea.maximumdirect.net/eric/narratio/internal/fileops"
+)
+
+// MaxExternalResultFileBytes bounds transcript and artifact files produced by
+// external adapters before a stage validates or materializes them.
+const MaxExternalResultFileBytes int64 = 64 * 1024 * 1024
+
+func readExternalResult(path, category string) ([]byte, error) {
+	data, err := fileops.ReadRegularFile(path, MaxExternalResultFileBytes)
+	if err != nil {
+		return nil, fmt.Errorf("%s exceeds or cannot be read within %d-byte limit: %w", category, MaxExternalResultFileBytes, err)
+	}
+	return data, nil
+}

internal/stage/extract.go

@@ -349,9 +349,9 @@ func checksumRegularFile(path string, requireJSON bool) (string, error) {
 	if info.Size() == 0 {
 		return "", fmt.Errorf("path %q must be non-empty", path)
 	}
-	data, err := os.ReadFile(path)
+	data, err := fileops.ReadRegularFile(path, artifacts.MaxExtractionPayloadBytes)
 	if err != nil {
-		return "", err
+		return "", fmt.Errorf("notarius lane result exceeds or cannot be read within %d-byte limit: %w", artifacts.MaxExtractionPayloadBytes, err)
 	}
 	if requireJSON && !json.Valid(data) {
 		return "", fmt.Errorf("path %q is not valid JSON", path)

internal/stage/polish.go

@@ -284,9 +284,9 @@ func validateProcessedTranscriptOutput(path string) error {
 	if err := requireFile(path, "processed transcript"); err != nil {
 		return err
 	}
-	data, err := os.ReadFile(path)
+	data, err := readExternalResult(path, "audita processed transcript result")
 	if err != nil {
-		return fmt.Errorf("read processed transcript: %w", err)
+		return err
 	}
 
 	var payload map[string]any

internal/stage/run_local.go

@@ -2,7 +2,6 @@ package stage
 
 import (
 	"fmt"
-	"os"
 	"path/filepath"
 	"strings"
 
@@ -130,9 +129,9 @@ func materializeRunLocalOutput(
 	if canonicalPath == "" {
 		return artifacts.Ref{}, fmt.Errorf("canonical destination path is required")
 	}
-	data, err := os.ReadFile(srcPath)
+	data, err := readExternalResult(srcPath, "run-local external output")
 	if err != nil {
-		return artifacts.Ref{}, fmt.Errorf("read run-local output %q: %w", srcPath, err)
+		return artifacts.Ref{}, err
 	}
 	if err := store.WriteFileAtomic(canonicalPath, data, fileops.WorkspaceFileMode); err != nil {
 		return artifacts.Ref{}, fmt.Errorf("materialize output to %q: %w", canonicalPath, err)

internal/stage/transcribe.go

@@ -321,9 +321,9 @@ func validateTranscriptJSONFile(path string) error {
 	if err := requireFile(path, "raw transcript"); err != nil {
 		return err
 	}
-	data, err := os.ReadFile(path)
+	data, err := readExternalResult(path, "whisperx transcript result")
 	if err != nil {
-		return fmt.Errorf("read transcript: %w", err)
+		return err
 	}
 	var payload any
 	if err := json.Unmarshal(data, &payload); err != nil {

internal/stage/trim.go

@@ -3,7 +3,6 @@ package stage
 import (
 	"context"
 	"fmt"
-	"os"
 	"path/filepath"
 	"strings"
 	"time"
@@ -389,9 +388,9 @@ func resolveTrimmedOutputPath(paths artifacts.SessionPaths, cfg *config.TrimConf
 }
 
 func copyTranscript(store artifacts.Store, src, dst string) error {
-	data, err := os.ReadFile(src)
+	data, err := readExternalResult(src, "seriatim trimmed transcript result")
 	if err != nil {
-		return fmt.Errorf("read source transcript: %w", err)
+		return err
 	}
 	if err := store.WriteFileAtomic(dst, data, fileops.WorkspaceFileMode); err != nil {
 		return fmt.Errorf("write destination transcript: %w", err)