thoughtproof-reasoning-check
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill performs network requests to api.thoughtproof.ai to verify reasoning claims. This involves sending potentially sensitive trade or logic details to an external domain.
- [INDIRECT_PROMPT_INJECTION]: The skill processes natural language reasoning (the 'claim' field) which is then analyzed by remote models. This creates a vulnerability surface where malicious instructions embedded in the reasoning could attempt to influence the verification outcome. 1. Ingestion points: The 'claim' field in the JSON payload sent to the /v1/check endpoint. 2. Boundary markers: None provided; the claim is sent directly as a string. 3. Capability inventory: The skill is explicitly designed to gate high-value autonomous actions, such as token swaps via the moonpay-swap-tokens skill. 4. Sanitization: There is no evidence of filtering or sanitizing the claim content before transmission to the reasoning models.
Audit Metadata