openfin-onramp

Pass

Audited by Gen Agent Trust Hub on May 12, 2026

Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [SAFE]: The skill follows security best practices by delegating sensitive payment details and KYC processes to trusted third-party providers (Moonpay and Onramp.money) and explicitly instructing the agent not to sign on-chain transactions or solicit credit card/payment information.
  • [PROMPT_INJECTION]: The skill defines an input surface where user-supplied data (currency, chain, amount) is used to construct API requests, constituting a surface for indirect prompt injection.
  • Ingestion points: User-provided parameters for the get_onramp_url tool and /agent/onramp endpoints.
  • Boundary markers: Absent.
  • Capability inventory: Network POST requests to internal API routing endpoints.
  • Sanitization: Not explicitly defined in the skill instructions.
  • [DATA_EXFILTRATION]: The skill performs network requests to vendor-managed endpoints (openfinance.tech) to retrieve checkout URLs. These operations are functional for the skill's purpose and do not involve unauthorized access to sensitive local files or credentials.
Audit Metadata
Risk Level
SAFE
Analyzed
May 12, 2026, 07:59 PM
Security Audit — agent-trust-hub — openfin-onramp