scholar-ideation

Pass

Audited by Gen Agent Trust Hub on Jul 17, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill performs legitimate academic research tasks using agent tools (WebSearch, WebFetch, Write) as intended for its primary purpose.
  • [SAFE]: No evidence of prompt injection, obfuscation, or unauthorized data exfiltration was detected in the skill instructions or associated phase documents.
  • [SAFE]: External communication is restricted to well-known academic databases (arXiv, Google Scholar) and user-configured MCP tools for Zotero integration.
  • [SAFE]: The skill incorporates human-in-the-loop checkpoints via AskUserQuestion, ensuring the user reviews and confirms research directions and keywords.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 17, 2026, 12:17 PM
Security Audit — agent-trust-hub — scholar-ideation