explore
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- SAFE (SAFE): The 'explore' skill follows its stated purpose of assisting with code search and navigation. Analysis across all 10 threat categories confirms the absence of malicious behavior. Specifically: no prompt injection patterns are present; no network operations or external data exfiltration are defined; no obfuscation techniques are used; and all referenced commands (find, grep, git, lsp) are standard local utilities for code analysis. Indirect prompt injection risk is considered negligible as the agent only processes local code files for search purposes within the intended user context.
Audit Metadata