# @thinkhive/mcp-explainer

Source: https://glimind.com/server/npm%2F-thinkhive%2Fmcp-explainer
Registry id: `npm/-thinkhive/mcp-explainer`
Status: unknown

## Install

```json
{
  "mcpServers": {
    "thinkhive-mcp-explainer": {
      "command": "npx",
      "args": [
        "-y",
        "@thinkhive/mcp-explainer"
      ]
    }
  }
}
```

Claude Code:

```sh
claude mcp add thinkhive-mcp-explainer -- npx -y @thinkhive/mcp-explainer
```

## Tools (25)

- `cluster_details`: Get cluster details with examples
- `cluster_list`: List failure clusters for an agent
- `eval_get_results`: Get evaluation results
- `eval_list_sets`: List evaluation sets
- `eval_run`: Run evaluation against golden dataset
- `fix_generate`: Generate fix suggestions for failures
- `fix_list`: List generated fixes
- `quality_get_scores`: Get quality scores for trace/agent
- `quality_hallucination_detect`: Detect hallucinations in responses
- `quality_rag_evaluate`: Evaluate RAG quality (groundedness, faithfulness)
- `roi_calculate_impact`: Calculate business impact for failures
- `roi_get_correlations`: Get performance-to-business correlations
- `roi_get_summary`: Get ROI summary with revenue impact
- `silent_failures_detect`: Detect silent failures
- `silent_failures_summary`: Get silent failure summary
- `trace_batch_analyze`: Analyze multiple traces at once
- `trace_explain`: Analyze with RAG eval & hallucination detection
- `trace_get_explanation`: Get stored explanation by trace ID
- `trace_patterns`: Get failure patterns for an agent
- `trace_search`: Search traces using natural language
- `trace_stats`: Get statistics for traces
- `webhook_create`: Create a new webhook
- `webhook_delete`: Delete a webhook
- `webhook_list`: List configured webhooks
- `webhook_test`: Test a webhook

Measured by Glimind (https://glimind.com): MCP handshake + tool list only, no tool is ever called. CC BY 4.0.