The Agent Skills Directory

[PROMPT_INJECTION]: The skill possesses an indirect prompt injection attack surface due to its interaction with external data providers. \n- Ingestion points: The skill fetches PR titles, states, and coverage details from GitHub and Codecov using the referenced script scripts/codecov_diff.py. \n- Boundary markers: No boundary markers or special instructions are present in the skill to prevent the agent from potentially following malicious instructions embedded in the external coverage data or pull request metadata. \n- Capability inventory: The skill is capable of executing shell commands and local Python scripts, and it utilizes the authenticated gh CLI. \n- Sanitization: Sanitization and filtering of external data cannot be verified as the content of the referenced scripts/codecov_diff.py is not provided in the skill package. \n- [COMMAND_EXECUTION]: The skill instructs the agent to run a local script scripts/codecov_diff.py using the uv package manager and relies on the pre-authenticated GitHub CLI to interact with repositories.

coverage