skills/security-phoenix-demo/security-skills-claude-code/production-readiness-review/Gen Agent Trust Hub
production-readiness-review
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a bundled script
scan_repo.shto perform repository analysis. This tool is restricted via theallowed-toolsconfiguration to only run the specific script provided with the skill, preventing arbitrary command execution. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from external repositories during Phase 1 (deterministic scan) and Phase 2 (requirement trace), creating a surface for indirect prompt injection.
- Ingestion points: File contents and directory structures read from the target repository, and the output of the
scan_repo.shscript. - Boundary markers: The instructions do not specify explicit delimiters or "ignore" instructions for the ingested code snippets.
- Capability inventory: The agent has restricted
Bashaccess (limited toscan_repo.sh) and instructions to modify files during Phase 3 (mutation testing) and Phase 6 (remediation). - Sanitization: The skill relies on model judgment and does not specify explicit sanitization steps for the ingested code content.
Audit Metadata