| Term | Definition |
|---|---|
| Decision | One agent turn: input + output (+ metadata), ingested for monitoring and evaluation |
| Agent profile | App entity defining an agent’s name, role, description, and analytics |
| agent_id | String identifier you send from code to label which agent produced a decision |
| session_id | Groups decisions in one conversation or workflow run |
| domain | Evaluation context string (e.g. customer_support) |
| Dimension | One of eight judgment axes scored 0–10 |
| Overall score | Aggregate judgment score derived from dimensions |
| Grade | Letter grade (A–F) from overall score |
| Warmup | Initial period before evaluation starts; requires minimum decision count |
| Incident | Evaluation flag when score drops below threshold |
| Streak | Repeated low scores triggering inbox attention |
| Ingest | API/SDK path that accepts decisions |
| Trace | SDK construct for multi-step agent workflows |
| Workspace | Your organization’s isolated data boundary |
| API key | Secret token for SDK/REST ingest authentication |
| Public profile | Shareable URL showing agent judgment analytics |
| Command center | Dashboard home screen in the app |
| Safe completion rate | Metric indicating decisions within policy boundaries |
Resources
Glossary
Terms used across Histeeria docs and the product.