beacon

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

• Third-party content exposure detected (high risk: 0.80). The skill's SKILL.md shows it issues Membrane proxy requests and action runs against the Beacon API (e.g., "membrane request CONNECTION_ID /path/to/endpoint" and action run commands) which ingest customer-generated Beacon data (user messages/notes) that the agent would read and could influence subsequent actions.