The Agent Skills Directory

[SAFE]: The skill's primary functions—analyzing writing styles (stylometry) and generating content based on those styles—are implemented using standard logic and instructions without any detected malicious patterns.
[DATA_EXPOSURE]: The skill instructions include steps to read local files and folders using Read and Glob tools with specific patterns (**/*.{md,txt}). This is a legitimate requirement for the skill to analyze a local text corpus as requested by the user.
[EXTERNAL_DOWNLOADS]: The skill identifies and fetches content from external URLs (Substack, blogs) to build its analysis corpus. This is a documented feature and utilizes standard web fetching tools (WebFetch) to retrieve public writing samples for stylometric analysis.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external URLs and local files (the corpus). While this constitutes an ingestion surface, the risk is mitigated by explicit instructions to the AI to analyze only the 'form' (DNA) of the writing and use a fixed markdown template for output. Boundary markers like --- are also suggested for content separation.

style-writer