idea-evaluator

Pass

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill follows the principle of least privilege by operating in a read-only mode. It explicitly instructs the agent not to run setup scripts (e.g., scripts/init-writing-dir.sh) or modify the user's .writing/ state.
  • [SAFE]: The skill implements robust integrity and anti-hallucination measures, such as the [UNVERIFIED] tagging for literature claims and the requirement to cite specific evidence for every score given.
  • [SAFE]: No network access, data exfiltration patterns, or external code execution vectors were found. All evaluation logic is self-contained within the provided Markdown references.
  • [SAFE]: The skill uses structured gates (Fatal Flaws, FINER, Integrity Gate) to ensure the AI's output is consistent and grounded in provided data, reducing the risk of erratic behavior.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 13, 2026, 05:55 PM
Security Audit — agent-trust-hub — idea-evaluator