deep-code-review
Pass
Audited by Gen Agent Trust Hub on Aug 25, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a secure-by-default code review workflow, mandating read-only access and isolated environments for any necessary code execution. It explicitly forbids the use of credentials or write-access tokens during the review process.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted code and documentation from repositories, which constitutes an attack surface. However, it provides robust mitigations:
- Ingestion points: Full repository content, including code, PR descriptions, and issues.
- Boundary markers: Section 2 of SKILL.md explicitly instructs the agent to treat instruction-like text encountered in these files as untrusted data rather than authority.
- Capability inventory: The agent is permitted to run deterministic tools (git, linters, tests) and shell commands.
- Sanitization: Section 6 requires the agent to inspect build scripts, hooks, and commands before execution to prevent malicious code from affecting the host environment.
- [SAFE]: The provided Python test suite (tests/test_report_contract.py) is benign, using standard libraries for internal consistency checks without external network or file system dependencies beyond the skill's own documentation.
- [SAFE]: No obfuscation, data exfiltration patterns, or persistence mechanisms were found in the skill's instructions or supporting configuration files.
Audit Metadata