extension-payment

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

• Third-party content exposure detected (high risk: 0.80). The skill's required workflow (Step 3) explicitly instructs the agent to use the docs-seeker skill to fetch live provider documentation and llms.txt files (e.g., https://docs.stripe.com/llms.txt and other provider docs listed in Step 2), which are public third-party pages the agent will read and act on to implement integrations, so those external pages could supply instructions that materially influence the agent's actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill instructs using the docs-seeker at runtime to fetch provider "llms.txt" files which are intended for AI-assisted docs and can inject remote prompt/instruction content (e.g. https://docs.stripe.com/llms.txt and https://docs.polar.sh/llms-full.txt), so those URLs directly control agent prompts during execution.

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

• Direct money access detected (high risk: 1.00). The skill is explicitly and specifically designed to integrate payment gateways and manage payments/subscriptions. It names payment providers (Stripe, PayPal, Paddle, Lemon Squeezy, Polar, etc.), requires API keys/product/price IDs and webhook URLs, instructs fetching provider API docs, and outlines the payment flow (payment page, webhook hits backend, backend validates/stores license). These are direct financial execution capabilities (payment gateway and subscription management), not generic tooling.