lov-fill-web-form
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill uses the
WebFetchtool in Step 1 to retrieve content from user-provided URLs. This content is then used to define the fields and context for a subsequent local knowledge base search. A malicious website could embed instructions in its form fields or page content designed to hijack the agent's behavior. - Ingestion points:
WebFetchcall in the 'Fetch and extract form fields' step (Step 1). - Boundary markers: Absent. The instructions do not specify using delimiters or 'ignore' commands when processing the fetched web content.
- Capability inventory:
WebFetch,Agent (Explore),Grep,Glob,Read, andWrite(all listed in compatibility section). - Sanitization: Absent. There is no mention of filtering or validating the structure of the retrieved form data before passing it to the sub-agent.
Audit Metadata