The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches a version manifest (manifest.json) from the author's GitHub repository to check for updates to the audit-rules skill.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it ingests and processes untrusted agent instruction files.
Ingestion points: The skill reads project files including AGENTS.md, CLAUDE.md, .cursor/rules/*.mdc, .windsurfrules, and several other agent-specific configuration files.
Boundary markers: There are no explicit boundary markers or instructions defined to prevent the agent from accidentally executing instructions found within the files it is auditing.
Capability inventory: The skill allows the agent to read file contents and perform targeted writes to the file system (following user approval).
Sanitization: Content from the audited files is not sanitized or escaped before being presented to the agent for diagnostic analysis.

audit-rules