The Agent Skills Directory

[SAFE]: The skill serves as a legitimate maintenance tool for documentation synchronization. It lacks network capabilities, does not access sensitive system files, and includes no obfuscated or persistent code.
[COMMAND_EXECUTION]: The skill leverages local command-line tools such as git, find, sed, jq, and comm to analyze repository structure and commit history. These operations are performed within the scope of the local project directory to maintain documentation accuracy.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the filesystem, including directory names and git commit messages.
Ingestion points: Reads file listings from skills/, .agents/skills/, and .claude/skills/, and commit subjects from git log.
Boundary markers: None explicitly present in the resulting documentation list.
Capability inventory: Limited to reading/writing documentation files and querying git metadata.
Sanitization: Employs jq --arg for safe interpolation of skill names into JSON queries, preventing command injection during classification logic. The overall risk is assessed as SAFE.

update-docs