fredyee/hallucc-mcp

Overview

MCP server for detecting AI hallucinations and enforcing agent safety. It verifies claims, guards prompts, and gates computer-use agent actions. Written in TypeScript.

Best for

Best for
Developers building MCP-based agents that need claim verification and safety gating

Use cases

• Verify model-generated claims against trusted sources
• Guard prompts against injection or unsafe instructions
• Gate CUA actions before they execute

How to use

Install

npx @modelcontextprotocol/inspector

Tools exposed

• verify_text
• verify_agent
• check_cua_actions
• check_cua_code_audit
• check_safety

Tested with

Claude Desktop, Claude Code, Cursor

Notes

MCP server for detecting AI hallucinations and enforcing agent safety. It verifies claims, guards prompts, and gates computer-use agent actions. Written in TypeScript.

0 stars on GitHub. Last updated 2026-09-23. Licensed MIT.

Use cases

• Verify model-generated claims against trusted sources
• Guard prompts against injection or unsafe instructions
• Gate CUA actions before they execute

Pros

• Covers both hallucination detection and agent safety in one MCP server
• Uses the MCP standard, so it can plug into MCP-compatible agent workflows
• TypeScript codebase for developers in the JS/TS ecosystem

Cons

• Zero stars and no adoption data, so maturity is unproven
• Limited to MCP-compatible environments
• No documentation or usage examples provided in the available facts

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.