sumup

Warn

Audited by Snyk on Nov 17, 2026

Risk Level: MEDIUM
Full Analysis

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

  • Third-party content exposure detected (low risk: 0.10). The required workflow runs Membrane actions and proxy requests that pull SumUp API records (transactions, refunds, payouts, merchant profile), which may contain payer-influenced fields such as descriptions or refund reasons, but the agent fetches these on demand and the skill defines no outsider submission path into a monitored queue.

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

  • Direct money access detected (medium risk: 0.30). Operation: The skill is a SumUp integration that exposes Payment, Transaction, Payout, and Refund objects (SKILL.md:23-27) and lets the agent run any discovered SumUp action or send raw requests to the SumUp API through a connection (SKILL.md:113-125, 130). Payments, refunds, and payouts are direct money movements, and delegated execution through this route counts. The skill is not merely a tutorial or development guide, so the financial-scope exclusions do not apply.

Type: Direct. Refunds, payouts, and payment transactions move money, so the indirect test does not apply.

Control: A qualifying control is documented. Access requires a Membrane login (SKILL.md:41-47). A SumUp connection must reach READY state, and when it is not ready the user must authenticate through OAuth or a provided UI (SKILL.md:85-93). The execution layer then injects the credentials for every action and proxy request on that connection (SKILL.md:130). This authenticated connection gates the SumUp routes, including the payment, refund, and payout routes. The gap is that the skill does not document per-action authorization or approval for money-moving actions. Because a service-enforced authentication gate is documented, the direct operation meets the Medium threshold rather than High.

Issues (2)

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
MEDIUM
Analyzed
Nov 17, 2026, 07:05 PM
Issues
2