leadoo

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

• Third-party content exposure detected (high risk: 0.70). The SKILL.md instructs the agent to connect to Leadoo via Membrane (e.g., "membrane connect --connectorKey leadoo") and run actions (e.g., "membrane action run ...") whose output returns visitor/conversation/lead data (user-generated) that the agent is expected to read and act on, so untrusted third-party content can influence subsequent decisions.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill requires and runs the Membrane CLI (installed/run via npm/npx) which fetches and executes remote code at runtime — e.g. the package at https://www.npmjs.com/package/@membranehq/cli — so this external dependency can execute code during skill operation.