solo-swarm
Pass
Audited by Gen Agent Trust Hub on Feb 21, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION] (LOW): Potential for Indirect Prompt Injection (Category 8) from processing untrusted data from web sources like Reddit and GitHub. Evidence Chain: 1. Ingestion points: WebSearch and WebFetch tool outputs. 2. Boundary markers: Absent. 3. Capability inventory: Bash, Write, and WebFetch tools. 4. Sanitization: Absent.
- [COMMAND_EXECUTION] (LOW): Uses the Bash tool to verify the status of a local search service (SearXNG) on localhost, which is a low-risk diagnostic action.
Audit Metadata