skills/openai/skills/gh-fix-ci/Gen Agent Trust Hub

gh-fix-ci

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • Command Execution: The skill uses subprocess.run to interact with git and the GitHub CLI (gh). This is a necessary component of its functionality to inspect repository state and retrieve logs. The implementation follows security best practices by passing command arguments as lists rather than using a shell environment, which helps prevent command injection.\n- Indirect Prompt Injection Surface: The skill ingests untrusted data in the form of GitHub Action logs to identify failure points.\n
  • Ingestion points: Logs are retrieved via gh run view --log and the GitHub API in scripts/inspect_pr_checks.py.\n
  • Boundary markers: The output does not currently use specific delimiters or instructions to treat log content as data only.\n
  • Capability inventory: The skill allows for local command execution and file writing to implement proposed fixes.\n
  • Sanitization: Log content is processed for snippets without explicit sanitization. This is a common characteristic of log-analysis tools and is considered a low-severity surface for the skill's primary purpose.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 01:17 PM
Security Audit — agent-trust-hub — gh-fix-ci