shipready
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill operates by ingesting and analyzing untrusted code from external repositories to perform audits.
- Ingestion points: The scripts scripts/recon.py, scripts/secret_scan.py, and scripts/env_check.py read file contents and metadata from target repositories.
- Boundary markers: The SKILL.md file includes extensive instructions for epistemic discipline, requiring the agent to cite repository facts and avoid external assumptions.
- Capability inventory: The skill can execute git commands via scripts/diff_helper.py to analyze project changes.
- Sanitization: Content is processed using regex and static parsing logic without executing the ingested data.
- [COMMAND_EXECUTION]: The helper script scripts/diff_helper.py executes local git commands for repository analysis.
- Evidence: The script uses subprocess.run to call git status and git diff. These calls use a list of arguments and do not invoke a shell, providing safety against command injection from user-supplied inputs.
Audit Metadata