github-actions-workflow-pro
Pass
Audited by Gen Agent Trust Hub on Nov 6, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill promotes and implements industry-standard security practices for CI/CD, such as requiring third-party actions to be pinned to immutable commit SHAs and favoring OIDC for cloud authentication over static secrets.
- [CREDENTIALS_UNSAFE]: The skill retrieves GitHub session tokens using the official
gh auth tokencommand. This process is managed securely by thescripts/scan.shwrapper, which injects tokens into child processes using environment variables, preventing sensitive data from appearing in execution logs, shell traces, or command-line arguments. - [COMMAND_EXECUTION]: The skill invokes several external scanners and local scripts via the shell. Technical review of
scripts/run-stats.pyconfirms thatsubprocess.runis called using argument lists rather than shell strings, effectively mitigating command injection risks. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from GitHub (PR titles and commit messages) via
scripts/run-stats.py. It provides a mandatory evidence chain of protection: untrusted content is sanitized and truncated in the script, and the instructions explicitly direct the agent to treat this information solely as data and ignore any embedded instructions. - [EXTERNAL_DOWNLOADS]: The skill references several well-known security tools (e.g., actionlint, zizmor) and the author's own repositories. These are trusted or well-known sources used for the skill's primary DevOps functions.
Audit Metadata