setup-asiapay

Fail

Audited by Snyk on Nov 16, 2026

Risk Level: CRITICAL
Full Analysis

CRITICAL E004: Prompt injection detected in skill instructions.

  • Potential prompt injection detected (medium risk: 0.30). The skill is a legitimate AsiaPay integration guide with official domains and contacts and no malicious intent, but it tells the agent to accept a datafeed when any comma-separated value in the hash matches, which weakens verification beyond what a single-hash check needs (documented as plugin behaviour for secret rotation).

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

  • Third-party content exposure detected (low risk: 0.10). The required workflow reads only first-party AsiaPay docs and the user's own code, but the datafeed POST (step 7, SKILL.md:26-27) can carry payer-entered free text such as Holder and echoed remark, which an outsider can influence through the hosted payment page, so I rate it low rather than no because the datafeed is pushed to the merchant's endpoint and the agent must actively read it to be exposed.

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

  • Direct money access detected (medium risk: 0.30). The skill is mainly about building and debugging AsiaPay (PayDollar) integrations, which is development guidance and excluded from financial scope. However, step 10 and the Critical Rules also direct the agent to run refunds and voids server-side through onlineApi.jsp (OnlineReversal) and orderApi.jsp (Void, RequestRefund). Those are direct financial operations: refunds, voids and captures that release or authorize money. The workflow is framed as integration code that handles refunds, not as the agent processing a customer's refund itself, but the skill documents the capability for refund/void/capture actions via the Merchant API and describes refund job handling, so the operational route is arguably present. Control: Merchant API authentication is merchantId + loginId + password in the POST body, issued by AsiaPay. This is service-validated API credentialing enforced by the gateway, which qualifies as independent access control under the evidence standard. The secureHash on payment requests is also enforced by the service, which drops requests without a valid hash. Because the direct route has a qualifying service-enforced control, the tier is medium.

Issues (3)

E004
CRITICAL

Prompt injection detected in skill instructions.

W011
MEDIUM

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

W009
MEDIUM

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

Audit Metadata
Risk Level
CRITICAL
Analyzed
Nov 16, 2026, 05:14 AM
Issues
3