ap2-a2a-extension

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

• Third-party content exposure detected (low risk: 0.10). The workflow fetches documentation from specific URLs and web-searches GitHub for reference implementations and samples.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.70). The skill instructs fetching specification and implementation documents from ap2-protocol.org at runtime to govern agent behavior, but lacks established official first-party provenance for that domain, making it an unverifiable external runtime dependency.

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

• Direct money access detected (high risk: 1.00). The skill documentation describes how to implement the AP2 protocol as an A2A (Agent-to-Agent) extension, which handles payment-specific task flows, mandates, and inter-agent payment communication (such as Intent Mandates, Checkout Mandates, and Payment Mandates). It establishes cryptographic signing, risk signals, and payment processing tasks between agents.