wechat-webhooks

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

• Third-party content exposure detected (high risk: 0.85). The required runtime workflow processes the incoming WeChat Pay webhook request body/headers (outsider-origin attacker-controlled text) via request.text() / express.raw(...).toString('utf8') and then JSON parses it into variables used by the handler, so that free-form outsider content is present in the LLM context if the agent/skill forwards runtime payloads for reasoning/logging.

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

• Direct money access detected (high risk: 1.00). This skill is explicitly for a payment gateway (WeChat Pay) webhook integration. It describes verifying WeChat Pay signatures, decrypting AEAD_AES_256_GCM ciphertext with a 32-byte APIv3 key, and handling payment/refund events (TRANSACTION.SUCCESS, REFUND.SUCCESS). Although it focuses on receiving and processing notifications rather than initiating payments, it is specifically designed around a payment provider and uses provider-specific cryptographic keys and events—meeting the criteria for a payment-gateway-related skill.