Loading ToolBench
MCP server for AI Agent Evaluation Service using Universal Workflow Engine. Delegates to UnifiedEvaluationService which processes agent results through Universal Workflow Engine and performs bias analysis.
Generate a variation of an existing assessment task
Get available assessment criteria and quality standards
Get a specific assessment task by ID
Get available assessment tasks with optional filtering
Get bias analysis results for a completed comparative assessment
Get MCP server health status
Hot-reload all assessment task files from disk without restarting
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 14 | <=2025-11-25 | v2 |
Submit single assessment result for validation and storage (NO bias analysis). This tool validates and stores individual evaluation responses without calculating bias scores. For bias analysis, use submit_comparative_assessment with both baseline and variation responses.
Submit paired baseline and variation assessment results for bias analysis