parseur

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

• Third-party content exposure detected (high risk: 0.70). The skill instructs the agent to connect to Parseur mailboxes and run Membrane actions (e.g., "membrane connect --connectorKey parseur" and "membrane action run ...") to retrieve documents/data (emails, PDFs) from Parseur, which are untrusted/user-generated third-party contents the agent will read and could influence subsequent actions.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill requires installing/running the Membrane CLI via commands like "npm install -g @membranehq/cli@latest" and "npx @membranehq/cli@latest", which will fetch and execute remote package code at runtime (package: https://www.npmjs.com/package/@membranehq/cli).