The Agent Skills Directory

[SAFE]: The skill performs legitimate documentation tasks within the local project scope. It reads project specifications and writes analysis drafts and architectural rules to the local documentation directory.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted project documentation to generate new rules and drafts. 1. Ingestion points: Reads docs/product/business-flow.md, docs/standards/tech-stack-rule.md, docs/standards/directory-rule.md, and docs/standards/tools-rule.md in Step 1 and Step 3.2. 2. Boundary markers: None present; the skill interpolates the file contents directly into the analysis prompt context. 3. Capability inventory: File read access, directory creation (mkdir -p), and file write access to the docs/standards/ directory. 4. Sanitization: None identified. The skill generates static markdown documentation which poses a low risk of execution or escalation.

standards-task