The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill analyzes external CI log data, which creates a surface for potential indirect prompt injection attacks.\n
Ingestion points: In SKILL.md, the commands gh pr checks and gh run view --log-failed are used to retrieve status and log data from external GitHub Actions runs.\n
Boundary markers: The skill includes a 'Codex publication boundary' in SKILL.md that requires authorization for external writes, providing a process-level boundary, although it lacks technical string delimiters for the ingested content.\n
Capability inventory: The agent is tasked with fixing code failures, which involves file editing and potentially executing git push to resolve CI issues.\n
Sanitization: No explicit instructions are provided to sanitize or filter the log output before it is processed by the AI.\n- [COMMAND_EXECUTION]: The skill utilizes the GitHub CLI (gh) to retrieve pull request and CI status information.\n
Evidence: Standard commands like gh pr view, gh pr checks, and gh run view are documented in SKILL.md to facilitate the CI fix-loop workflow.\n
Context: The MIRROR.md file indicates the skill was originally sourced from the Cursor plugins repository on GitHub.

cursor-loop-on-ci