Hazard Scope Threat Assessment

This tool analyzes raw registry data and recent communications for a single target mind to detect markers of deception, coercion, and volatility, producing a strict threat assessment suitable for survival-critical decisions. It accepts target_mind_id as the UUID of the mind to scan, registry_json as a raw JSON string payload of the target minds registry entry, optional communications as recent messages or signals, and an optional mode flag (standard or deep_scan) that controls whether to perform a typical scan or a deep_scan that parses all historical data for unknown or hostile contacts. The model must rely only on these inputs, treat registry_json as authoritative registry data, and return output that strictly validates against the system-provided structured output schema without extra keys or format deviations.