The Agent Skills Directory

[PROMPT_INJECTION]: Indirect Prompt Injection Surface.
Ingestion points: The skill's primary function involves fetching GitHub Actions logs and job logs through the gh CLI. These logs often contain output from tests, build scripts, and dependencies that may be controlled or influenced by external contributors in pull requests.
Boundary markers: There are no explicit delimiters or system instructions used when presenting log snippets to the agent to prevent it from interpreting data within the logs as instructions.
Capability inventory: The skill is designed with high autonomy, instructing the agent to 'Fix proactively (no need for users approval), commit and push'. This allows the agent to modify local files and update remote repositories based on its analysis of the ingested data.
Sanitization: The skill does not perform sanitization, filtering, or validation of the log content before it is processed by the agent.
[COMMAND_EXECUTION]: The skill uses the gh and git command-line tools via Python's subprocess.run to inspect PR status and commit fixes. While the implementation correctly uses argument lists to prevent shell injection, the high degree of automation (pushing code changes without user review) increases the potential impact if the agent's logic is manipulated by untrusted log data.

gh-fix-ci