loogle-search
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute local commands including
loogle-search,loogle-server, and the Lean build toollake. These commands are executed against a local directory structure (~/tools/loogle). - [INDIRECT_PROMPT_INJECTION]: The skill facilitates the processing of user-supplied search patterns (e.g., "Nontrivial _ ↔ _") that are passed directly as arguments to shell commands.
- Ingestion points: Search patterns provided by the user in the proof or search context are ingested in
SKILL.mdexamples. - Boundary markers: The instructions lack explicit boundary markers or warnings to the agent to treat search patterns as untrusted data.
- Capability inventory: The skill uses
loogle-searchandloogle-serverexecution capabilities. - Sanitization: There is no mention of input validation or escaping for the search patterns before they are passed to the shell.
Audit Metadata