linkfox-tsearch-search
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONPERSISTENCE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill fetches and summarizes untrusted data from the web using the
scripts/tsearch_web_search.pyscript, creating a surface for injection attacks. - Ingestion points: Web search results, including page content, are retrieved via the LinkFox tool gateway API.
- Boundary markers: No explicit delimiters or instructions are provided to the agent to treat the fetched web content as untrusted or to ignore embedded instructions.
- Capability inventory: The script
scripts/tsearch_web_search.pywrites the full, untrusted API response to local files in thelinkfox/directory. - Sanitization: No sanitization or filtering of the external content is performed before it is processed by the agent.
- [PERSISTENCE]: The
references/onboarding.mddocumentation contains instructions for the agent to guide users in modifying their shell profile files (such as~/.bashrcand~/.zshrc) to store theLINKFOX_AGENT_API_KEY. This is a persistence mechanism for credentials that ensures they are active in all future shell sessions.
Audit Metadata