> ## Documentation Index
> Fetch the complete documentation index at: https://docs.medera.info/llms.txt
> Use this file to discover all available pages before exploring further.
# Crisis Response Agent
> Immediate, documented response to crisis language — no AI gating.
**Category:** Patient-Facing · **Signals:** Real-time text/voice · C-SSRS criteria · Practice escalation tree · On-call availability
Always-on across every patient touchpoint. Detect, escalate, document — never decide.
## Configuration
| Field | Description |
| :---------------- | :------------------------------------------ |
| `criteria` | C-SSRS-aligned crisis-language criteria |
| `escalation_tree` | Practice escalation tree + on-call schedule |
| `documentation` | Documentation requirements |
| `safety_policy` | Patient safety policy |
## Execution
1. Monitors real-time text and voice for crisis language
2. Classifies risk per C-SSRS-aligned criteria
3. Triggers the practice escalation tree without agent gating
4. Connects a human within seconds
5. Logs the event and the response for audit
## Composable experts
| Expert | Role |
| :-------------------------- | :------------------------------ |
| **Crisis Detection Expert** | C-SSRS-aligned criteria |
| **Escalation Expert** | Practice tree + on-call routing |
| **Documentation Expert** | Event logging + follow-up |
| **Safety Policy Expert** | Configurable safety guardrails |
## Compliance guarantees
* Agent gating is NEVER allowed in the crisis path
* Detection criteria are documented and configurable
* Every crisis event is logged with the response and the follow-up
## Endpoints
Canvas: `guardrails` (suicide-language rule) + `escalate`.
Care Orchestrator: `INTAKE_RED_FLAG_DETECTED` event.
***
## Related
Framework + 24 agents.
Typed expert registry.
Compose your own workflow.
REST surface.