security-sentinel
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION] (SAFE): The skill instructs the agent to use standard
grepcommands to search for patterns in local project files. These operations are restricted to searching and do not involve risky flags, piping to destructive shells, or unauthorized file modifications. - [PROMPT_INJECTION] (SAFE): The skill possesses an ingestion surface for untrusted data (Category 8). 1. Ingestion points: Local code files via grep. 2. Boundary markers: Absent. 3. Capability inventory: Local command execution (grep). 4. Sanitization: Absent. This is an inherent risk for an auditing tool and is categorized as SAFE in this context.
Audit Metadata