The Agent Skills Directory

[SAFE]: The skill performs legitimate novel project management tasks such as directory structure creation, character metadata migration, and project tracking initialization. All filesystem operations are scoped to the local writing project.
[SAFE]: The utility scripts (tracking_commit.py, build_outline_brief.py, check-outline-contract.js) implement robust validation logic, including filename sanitization (safe_file_component) and structured field extraction, which mitigates risks associated with directory traversal or malformed data input.
[INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided novel content (untrusted data) to perform deep analysis and generate outlines. The attack surface is managed through structured parsing logic that targets specific metadata (e.g., character names, plot points) rather than executing arbitrary text. This is a functional requirement of the skill and is handled with appropriate safeguards.

story-import