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