deep-research

Pass

Audited by Gen Agent Trust Hub on Mar 29, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill uses built-in orchestration patterns to manage research tasks. It utilizes native WebSearch and WebFetch tools for their intended purposes.
  • [SAFE]: The workflow includes explicit requirements for evidence triangulation, cross-checking claims across multiple sources, and mandatory citations, which promotes accurate and verifiable output.
  • [SAFE]: No sensitive file access, network exfiltration, or hardcoded credentials were detected in the skill's instructions or metadata.
  • [SAFE]: The instructions for parallel sub-agent execution (ResearchScout) are logical and scope-limited, focusing only on information retrieval and summarization.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 29, 2026, 01:18 AM