efficient-web-research
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill establishes a protocol for the agent to fetch and process data from untrusted external sources, including arbitrary websites and public GitHub repositories, creating a surface for indirect prompt injection.
- Ingestion points: The skill utilizes tools such as
read_url_content,browser_subagent, andsearch_webto retrieve content from the open web (SKILL.md). - Boundary markers: The instructions lack explicit requirements for the agent to use delimiters or to ignore potential instructions embedded within the fetched data.
- Capability inventory: The agent is equipped with web browsing, content reading, and search capabilities to execute the protocol.
- Sanitization: The protocol recommends stripping boilerplate and summarizing content, but these steps are focused on token efficiency rather than preventing the execution of instructions embedded in the external content.
Audit Metadata