adaptation
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill relies on a 'Story' command-line interface and external tools like
inklecateandTweegoto perform project maintenance and generate adaptation files. Fallback instructions allow the use of a local script located at../story-maintenance/scripts/story.jsvia Node.js or Bun.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-provided story content, including character names, locations, and glossary terms, which are interpolated into adaptation documents and CLI arguments.\n - Ingestion points: Content is read from
story.md, chapters, scenes, and various metadata files within the story project.\n - Capability inventory: The skill performs file reads and writes within the project structure and executes shell commands for building and validating content.\n
- Sanitization: The documentation notes that the build process for interactive formats escapes specific syntax in prose, reducing the risk of accidental code execution within those specific tools.\n
- Boundary markers: The skill follows a strict project schema, keeping adaptations in a dedicated
adaptations/directory to separate them from the source manuscript.
Audit Metadata