managing-github-ci
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONCREDENTIALS_UNSAFECOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references and provides templates using well-known GitHub Actions from trusted sources including
actions/checkout,actions/setup-node,pnpm/action-setup,changesets/action,peter-evans/create-pull-request, anddorny/paths-filter. These are standard tools in the DevOps ecosystem.\n- [INDIRECT_PROMPT_INJECTION]: The skill involves processing external data by instructing the agent to read and analyze GitHub Action workflow logs via theghCLI (e.g.,gh run view --log-failed). While these logs are ingested without explicit boundary markers or sanitization, the capabilities used (bash and file editing) are restricted to the primary task of troubleshooting CI failures, and the skill encourages local reproduction of errors as a verification step.\n- [CREDENTIALS_UNSAFE]: The documentation identifies the use ofGITHUB_TOKENandNPM_TOKENfor automation tasks. It adheres to security best practices by explicitly instructing users to configure these as repository secrets in GitHub Actions settings rather than hardcoding them or using insecure storage.\n- [COMMAND_EXECUTION]: The skill provides instructions for executing standard development commands usingpnpm,npm,npx, and the GitHub CLI (gh). These commands are intended for routine project management tasks such as testing, linting, building, and interacting with GitHub workflow runs.
Audit Metadata