The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes standard development tools including git, gh, and project-specific commands like npm run dev or rails server. These operations are essential for its function as a QA automation tool.
[EXTERNAL_DOWNLOADS]: A helper script (scripts/packs-resolve.py) manages the download of external 'Compound Packs' via git clone. The implementation follows security best practices, including the use of --end-of-options to mitigate command injection risks from external configuration.
[DYNAMIC_EXECUTION]: The skill uses Python's subprocess module to manage git operations and environment setup. The script performs careful ownership and directory boundary checks to ensure external content cannot access unauthorized parts of the file system.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external packs. It includes a specific mitigation instruction in references/phases.md telling the agent to treat this text as data to be recorded and quoted, rather than instructions to be followed.

ce-dogfood