livechat

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

• Third-party content exposure detected (high risk: 0.80). This skill's SKILL.md instructs the agent to connect to LiveChat via the Membrane CLI and run actions to access Conversations/Messages (user-generated chat content), which the agent will read and could use to drive responses or other actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill invokes the Membrane CLI via npx (npx @membranehq/cli@latest, which pulls code from the npm registry such as https://registry.npmjs.org/@membranehq/cli) at runtime, causing remote code to be fetched and executed and returning action definitions from the Membrane service that directly influence agent behavior.