certaindata

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt explicitly instructs the agent to insert the resolved API key value directly into outgoing Authorization headers when the env-var fast path is not available (and on platforms without substitution), which requires the LLM to include a secret verbatim in its generated output.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). SKILL.md ingests free text from the user to perform Bazaar semantic search and then parses attacker-influenced Bazaar results (resource.url / accepts[].outputSchema) before proceeding to payment or handoff via the Arbitrary x402 flow.

MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

• Direct money access detected (high risk: 1.00). The skill explicitly implements x402 payment signing and settlement: it (in "certaindata_flow" mode) calls a CertainData sign endpoint to obtain a payment signature, can fund a test wallet, retries the seller endpoint with the payment header, and captures on-chain settlement details (transaction hash, network). It orchestrates USDC-on-Base payments (production and sandbox), i.e. delegated signing and completing crypto payments — this is direct financial execution authority.