rw-claim-audit
Pass
Audited by Gen Agent Trust Hub on Jul 20, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a bundled Python helper script (
scripts/claim_audit.py) to initialize audit logs, add claim entries, and validate the audit state. This script is limited to standard library operations, performing local file I/O for state management and SHA256 hashing for document integrity verification. - [DATA_EXPOSURE]: The workflow involves reading manuscripts and cited references to verify information. These operations are performed locally within the agent's context to facilitate claim-to-source comparison. No network exfiltration or unauthorized file access was identified.
- [SAFE]: The skill includes robust internal logic such as behavior tests (
references/behavior-tests.json), methodology guides (references/method.md), and validation rules that emphasize manual verification over automated assumptions.
Audit Metadata