The Agent Skills Directory

[SAFE]: The skill defines a repeatable process for documentation maintenance, including research and reporting phases. It focuses on ensuring documentation quality and consistency with authoritative sources such as code definitions and API specifications.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it is designed to ingest and process documentation files from the repository.
Ingestion points: Markdown documentation across docs/, README.md, demos/, and governance files (CONTRIBUTING.md, etc.).
Boundary markers: None explicitly defined in the instructions.
Capability inventory: File reading across the repository, local command execution (make qualify), and potential file modification if requested by the user.
Sanitization: No specific sanitization or filtering of documentation content is mentioned.
[COMMAND_EXECUTION]: The skill mentions executing standard project-specific commands such as make qualify and the lychee link checker. These are routine development operations and are restricted to the local project environment.

aicr-auditing-docs