workflow-ship

Pass

Audited by Gen Agent Trust Hub on Jul 29, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill automates workflows by executing several shell commands, including git for version control and gh for GitHub repository management. It also invokes codex, a code review utility, assuming it is available on the system path.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by incorporating external data—such as branch names, git logs, and review tool outputs—directly into generated PR titles and summaries.
  • Ingestion points: Branch names (git branch --show-current), commit messages (git log), and review findings (codex review).
  • Boundary markers: None identified.
  • Capability inventory: Branch deletion (git branch -d), code pushing (git push), and pull request management (gh pr create/merge).
  • Sanitization: None identified; data is interpolated into command arguments for PR creation.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 29, 2026, 06:08 AM
Security Audit — agent-trust-hub — workflow-ship