you-web-search
Pass
Audited by Gen Agent Trust Hub on Aug 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local scripts such as
./secretcurland./notifyto perform authenticated API calls and send search results to the user. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted data from the web via search results and interpolates them into the agent's context and memory logs.
- Ingestion points: Data is ingested from the You.com Search API response processed in SKILL.md.
- Boundary markers: Absent; search snippets are placed directly into Markdown lists without delimiters.
- Capability inventory: The skill can perform network requests, write to the filesystem (logs and /tmp), and execute local helper scripts.
- Sanitization: Absent; no filtering is applied to the content of search snippets to remove potential instructions.
Audit Metadata