The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses node to execute local utility scripts (novel-outline.mjs, selftest.mjs). These scripts are part of the skill's distribution and perform data processing tasks such as text chunking, schema validation, and HTML rendering.
[EXTERNAL_DOWNLOADS]: The skill explicitly states it has 'zero dependencies' and 'zero API keys'. The renderHtml function produces a self-contained report with inline SVG/CSS and no external resource loads, ensuring privacy and offline functionality.
[DATA_EXFILTRATION]: No network exfiltration patterns were detected. The skill operates on local files provided by the user and outputs to local JSON, Markdown, and HTML files. Whitelisted domains or known paste sites are not present in the code.
[PROMPT_INJECTION]: The skill provides structured instructions for the AI agent to follow a multi-step adaptation workflow. There are no patterns of system prompt extraction or safety filter bypasses.
[DYNAMIC_EXECUTION]: While the skill involves script execution, the logic is static and deterministic. The novel-outline.mjs script acts as a validator and aggregator for the AI's output, enforcing strict rules (e.g., character counts, beat intervals) through code rather than LLM intuition.
[SAFE]: All analyzed scripts use standard Node.js built-in modules (node:fs, node:path, node:url) and do not perform suspicious operations. The self-test suite (selftest.mjs) is comprehensive and used for local verification of the adaptation logic.

novel-outline