genpark-agent-hallucination-fact-verifier-skill
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
genpark-agent-hallucination-fact-verifier-skill
🌐 GenPark MCP Hub • 📦 GenPark Official • 📖 Documentation
📌 Overview & Capability
genpark-agent-hallucination-fact-verifier-skill is a deterministic, high-performance, zero-dependency Python tool and native Model Context Protocol (MCP) server designed for autonomous AI agents, multi-agent frameworks (LangGraph, CrewAI, AutoGen, OpenAI Swarm), and developer environments (Cursor, Windsurf, Claude Desktop).
Executive Capability: Atomic claim decomposition and cross-context fact verification engine detecting ungrounded claims, numerical contradictions, and hallucinated entities against reference knowledge.
⚡ Key Highlights
🐍 Zero External
pipDependencies: Implemented entirely with pure Python standard library for instant zero-overhead execution.🔌 Native Model Context Protocol (MCP): Plugs directly into any MCP-compliant client via JSON-RPC 2.0 stdio.
⚡ Sub-Millisecond Execution: Slashes token burn and latency by resolving routine agent tasks deterministically without frontier LLM round-trips.
🛡️ Production-Hardened: Comprehensive error handling, boundary validation, and telemetry.
Related MCP server: mcp-citation-checker
🏗️ Architecture
graph LR
Agent([🤖 Autonomous Agent / IDE]) -->|MCP Protocol / JSON-RPC| Server[⚡ genpark-agent-hallucination-fact-verifier-skill Server]
Server --> Core[🧠 Deterministic Processing Core]
Core --> Out[📊 Actionable Result & Telemetry]
Out --> Agent🚀 Quickstart & Usage
1. Direct Python Client Execution
python example_usage.py2. Programmatic Integration
from client import AgentHallucinationFactVerifier
client = AgentHallucinationFactVerifier()
result = client.run_verification_benchmark()
print(result)🔌 Model Context Protocol (MCP) Setup
Connect this skill to Claude Desktop, Cursor, or any MCP-compliant client:
claude_desktop_config.json
{
"mcpServers": {
"genpark-agent-hallucination-fact-verifier-skill": {
"command": "python",
"args": ["/path/to/genpark-agent-hallucination-fact-verifier-skill/mcp_server.py"]
}
}
}Direct MCP Testing
python mcp_server.py --test📊 Technical Specifications
Parameter | Type | Required | Description |
|
| Yes | Primary context, code, schema, or content input |
|
| No | Execution flags, compression ratios, or risk bounds |
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Related MCP Connectors
Cross-check a factual claim against a verified knowledge graph before you assert it. Never guesses.
One-stop claim verification for AI agents: verdicts with confidence, evidence, signed receipts.
Deterministic verifier: LLM proposes, graph verifies. Sub-ms, offline, 0 tokens.
Deterministic claim verification with receipts across ~60 domains. No model in the loop.
Related MCP Servers
- AlicenseAqualityBmaintenanceMCP server for verifying AI agent claims vs reality — single-transcript inline grounding-check that flags when an agent's response states facts not in the input context, when its code silently swallows exceptions and substitutes mock data, or when its multi-turn transcript contains contradictions or unverified completion claims. Sub-second, local, free, no API calls.429 PyPI1MIT
- AlicenseNot gradedqualityDmaintenanceAn MCP server that verifies whether a claim is actually supported by the source text at a given citation — independent of what the calling LLM asserts.MIT
- AlicenseAqualityBmaintenanceMCP server for grounding verification (exact quotes, citations, code output, repo patterns, arithmetic) using no LLM in the verification path.52MIT
- AlicenseNot gradedqualityBmaintenanceEnables AI agents to cross-verify candidate claims against caller-supplied source texts, flagging hallucinations, numerical drift, entity mismatches, contradictions, and unverified assertions. It returns sentence-level verdicts with matched evidence snippets and machine-readable factual grounding confidence scores, exposed over MCP stdio, HTTP REST, and A2A discovery routes.1MIT