smart-explore
Pass
Audited by Gen Agent Trust Hub on Apr 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists of documentation that guides the agent on how to use AST-based search tools (smart_search, smart_outline, smart_unfold) to navigate codebases more efficiently by reducing token consumption compared to standard file-reading methods.
- [PROMPT_INJECTION]: The skill contains directives such as 'This skill overrides your default exploration behavior' and 'Your next action should be one of...'. These instructions are interpreted as operational guidelines for optimizing agent performance in a specific task domain and do not attempt to subvert safety filters or general instructions.
- [DATA_EXFILTRATION]: There are no indicators of data harvesting or external transmission; the skill does not utilize network-capable commands or target sensitive system paths like credentials or SSH keys.
- [COMMAND_EXECUTION]: While the skill defines interactions with local tools via MCP, it does not include execution of arbitrary shell commands, scripts, or privilege escalation attempts.
Audit Metadata