The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to retrieve and process content from external GitHub resources, including issue bodies, pull request descriptions, and workflow failure logs (e.g., in references/gh/SKILL.md and github_project_setup.py). This content is attacker-controllable and could harbor malicious instructions intended to manipulate the agent. Ingestion points include commands like gh pr view, gh issue view, and gh run view --log-failed. The provided instructions do not include boundary markers or explicit guidelines for the agent to treat this retrieved content as untrusted. The skill's capability inventory includes shell command execution and API-based modification of GitHub resources. No sanitization or filtering of the external text is implemented in the provided scripts.\n- [COMMAND_EXECUTION]: The script references/gh/scripts/github_project_setup.py utilizes the subprocess.run function to execute GitHub CLI (gh) commands for GraphQL interactions and project management. This is a required functional component for the skill's intended purpose.\n- [DYNAMIC_EXECUTION]: The test suite references/gh/tests/test_github_project_setup.py uses the importlib.util module to dynamically load and execute the github_project_setup.py script. Although this is a standard practice for testing Python scripts located in non-package directories, it involves the dynamic execution of local code at runtime.

github-workflows