doc
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local script
scripts/create_doc.pyto interact with Google Drive and Docs. This is a functional requirement for the skill's stated purpose of document creation and does not involve arbitrary or remote command execution. - [PROMPT_INJECTION]: The skill processes user-supplied and potentially scraped content, creating an indirect prompt injection surface.
- Ingestion points: User-provided content for document blocks and external scraped content as mentioned in the documentation.
- Boundary markers: The skill includes a prominent block at the top of SKILL.md instructing the agent to treat scraped content as data, not instructions, and to stop if it addresses the agent directly.
- Capability inventory: The skill has the capability to create and update files in Google Drive via the
create_doc.pyscript. - Sanitization: Content is parsed and structured into a specific JSON block schema (TITLE, BODY, BULLET, etc.) before being passed to the rendering engine, which limits the potential for instruction leakage.
Audit Metadata