regulatory-threat-model

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

• Third-party content exposure detected (medium risk: 0.30). The required runtime workflow ingests outsider-authored free text via the user’s system-description/key-assets intake (including repository-derived reconstruction) and then uses that data to drive workflow steps like start_workflow and subsequent tool interactions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill requires the Ansvar Gateway MCP connector at https://gateway.ansvar.eu/mcp which is called at runtime to return workflow steps and control the agent's actions (start_workflow / get_current_step / submit_response), so this external endpoint directly steers prompts and is a required runtime dependency.