mission-control

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

• Third-party content exposure detected (high risk: 1.00). The skill routinely fetches and processes user-generated content from the Mission Control API (e.g., heartbeat notifications including soul_md_sync content, /api/tasks titles/descriptions/comments, /api/squad-chat broadcasts and direct-messages, and /api/documents), which are untrusted third-party inputs the agent is expected to read and act on and therefore could enable indirect prompt injection.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill calls the Mission Control API (default https://missioncontrol.ai, configurable via MISSION_CONTROL_API_URL) on every heartbeat and will, if soul_md_sync.required is true, write the returned soul_md_sync.content into the agent's SOUL.md at runtime—allowing that external URL to directly modify the agent's instructions/prompts.