security-review
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFENO_CODECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is entirely instructional and focused on security best practices, explicitly forbidding the exfiltration of secrets.
- [NO_CODE]: No scripts or binaries are included, limiting the functionality to model-driven analysis.
- [COMMAND_EXECUTION]: The workflow involves running git commands to view code changes, which is standard behavior for development-related agents.
- [PROMPT_INJECTION]: While the skill processes untrusted code via git diff, which is an indirect prompt injection surface, its primary purpose is to identify and mitigate such risks through multi-pass auditing. 1. Ingestion points: Git diff output in SKILL.md. 2. Boundary markers: Not explicitly defined. 3. Capability inventory: Spawning subagents and reading local files. 4. Sanitization: Absent.
Audit Metadata