The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources that could contain malicious instructions designed to manipulate the agent. 1. Ingestion points: Reads content from gh pr checks output, GitHub Actions logs, and follows links to external check services (SKILL.md). 2. Boundary markers: The instructions do not define markers or specific directives to isolate untrusted logs from the agent's system prompt or following instructions. 3. Capability inventory: The skill has the capability to modify repository files and push code changes to the branch (SKILL.md). 4. Sanitization: No sanitization or validation of the fetched log content or link data is specified.
[COMMAND_EXECUTION]: The skill uses the GitHub CLI and git commands to perform its duties. Evidence: The workflow utilizes gh pr checks and performs push operations after applying fixes.

fix-ci