The Agent Skills Directory

[COMMAND_EXECUTION] (LOW): The skill executes local gh and git commands to diagnose and fix repository issues. This includes write-access operations like git push to a fix branch and gh run rerun.
[PROMPT_INJECTION] (LOW): The skill is vulnerable to Indirect Prompt Injection (Category 8) due to its data processing workflow. An attacker could potentially influence the agent's code fixes by injecting instructions into failing test logs or artifacts. Ingestion points: Failure logs from gh run view --log-failed and downloaded artifacts from gh run download. Boundary markers: None; the agent is prompted to analyze raw log content for root causes. Capability inventory: git add, git commit, git push, and workflow modification capabilities. Sanitization: No validation or sanitization of log content is performed before the agent implements and pushes a fix.

ci-fix