← DocsAI Risk Scoring v2.1.0
⚠️ Assistive scoring only. Not legal advice. Not compliance guarantee.
1. Routing
| Confidence | Provider | data_handling |
|---|
| ≥70% | Ollama | local_only |
| <70% | DeepSeek | external_processor |
2. Response Schema
{"risk_score":75,"risk_level":"HIGH","confidence":85,"key_risks":["Anonymous team"],"compliance_flags":["Unregistered"],"recommendation":"Require audit","from_cache":false,"source_provider":"deepseek","source_model":"deepseek-chat","model_version":"v3","pipeline_version":"2.1.0","prompt_version":"risk_v3","doc_hash":"sha256:...","cache_key":"sha256:...","data_handling":"external_processor","created_at":"2026-01-18T12:00:00Z","request_id":"req_123","validated":true}
3. Risk Levels
| Score | Level |
|---|
| 0-29 | LOW |
| 30-59 | MEDIUM |
| 60-79 | HIGH |
| 80-100 | CRITICAL |
4. Request
POST /api/v1/risk/analyze
{"text":"...","allow_external":true}
5. Caching
TTL 30d. cache_key=sha256(doc_hash+prompt_version+model+pipeline_version)
6. Validation
Schema validation + repair step for invalid LLM output.
7. ChatGPT
Import: https://feedoracle.io/.well-known/openapi.json