gh-fix-ci
Pass
Audited by Gen Agent Trust Hub on Apr 18, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill uses
gitandgh(GitHub CLI) to retrieve PR logs and status information. Analysis ofscripts/executable_inspect_pr_checks.pyconfirms that commands are executed using argument lists viasubprocess.run, which prevents shell injection vulnerabilities.- [PROMPT_INJECTION]: The skill processes untrusted GitHub Action logs, which presents a surface for indirect prompt injection. This risk is appropriately mitigated by the skill's design, which requires explicit user approval before the agent implements any suggested fixes. - Ingestion points: PR logs are fetched in
scripts/executable_inspect_pr_checks.py(lines 218 and 231). - Boundary markers: The logs are processed as raw text windows without specific delimiters for the LLM.
- Capability inventory: The skill allows the agent to propose and apply code changes after user confirmation.
- Sanitization: Logs are trimmed for context but are not filtered for malicious instructions.- [SAFE]: No indicators of data exfiltration, credential harvesting, or malicious persistence were found in the scripts or instructions.
Audit Metadata