Documented as a crisis-detection and escalation-routing system built around mandatory human-in-the-loop review for critical cases, not autonomous action.
The README documents a 4-layer security pipeline (input validation, rate limiting, abuse detection, cryptographic signing), a 5-tier risk classifier, and a cryptographic Merkle-tree audit chain — with an explicit boundary that critical cases are never escalated autonomously; a human reviewer is required.
mental-health-escalation-router 0.1.0
Source: pyproject metadata at pinned commit.
Open metadata sourceDespite production-grade marketing language in the README, the repository's own status badge marks it Development, not production — this portal records the system as Partial on that basis rather than upgrading the self-description.
This portal did not independently test the claimed <1000ms latency bound, the 5-tier risk classifier's accuracy, or the 400+ test suite referenced in the README.