google-docs-manager
Pass
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to an indirect prompt injection/query injection attack in the document search functionality.
- Ingestion points: The
queryparameter in thefindcommand withinscripts/docs.pyis directly accepted from user input. - Boundary markers: Absent; the input is interpolated directly into the API query string without delimiters or warnings.
- Capability inventory: The script possesses capabilities to list, read, modify, and delete Google Docs (
docs.py). - Sanitization: The
cmd_findfunction inscripts/docs.pylacks proper escaping for single quotes when constructing the search string:query = f"... name contains '{args.query}' and trashed=false". An attacker could provide a crafted query (e.g., "' or 1=1 or name contains '") to bypass intended search filters and discover documents they should not see.
Audit Metadata