The Agent Skills Directory

[SAFE]: No malicious patterns or security vulnerabilities were detected. The skill is designed with a strong defensive posture, explicitly forbidding the creation of empty directory structures or the writing of draft content to canonical paths. It implements a human-in-the-loop workflow where every documentation proposal must pass a structural validator and receive explicit user approval before any file writes occur.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from codebase files and manifests to generate documentation evidence. The design incorporates a mandatory evidence chain requiring human confirmation for facts and structural validation for all proposed changes, which prevents the automatic execution of instructions that might be embedded in the processed data.

define-system-doc