docs
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided inputs, such as document titles, memos, and pasted artifact links (e.g., from claude.ai), which are then used to parameterize tool calls. Specifically, the agent is instructed to call a
guidetool withtopic.instructions, creating a surface where malicious instructions embedded in a user-provided topic or external link could influence the agent's logic. - Ingestion points: User requests naming a document, page, or PRD; pasted artifact links.
- Boundary markers: The skill lacks explicit instructions or delimiters to isolate user-provided data from tool execution logic.
- Capability inventory: The skill enables document creation, file reading, and searching.
- Sanitization: There are no mentioned procedures for sanitizing or validating the
topicorinstructionsparameters before they are used by the docs connector.
Audit Metadata