coherence
MCP server that surfaces coherence, governance, and safety primitives: Wave/Bump validation, ATOM trail + gates, .context.yaml packing, AWI intent scaffolding, and docs/search across the SpiralSafe corpus.
claude mcp add --transport stdio toolate28-coherence-mcp npx -y @toolate28/coherence-mcp \ --env ATOM_AUTH_TOKEN="Required for authenticated operations (set your token)" \ --env WAVE_TOOLKIT_BIN="Optional path to wave-toolkit CLI" \ --env SPIRALSAFE_API_TOKEN="Required for ops tools (set your token)"
How to use
coherence-mcp is a dedicated MCP server that validates coherence across AI systems using the Tri-Weavon architecture. It exposes a suite of 49 tools for tasks like WAVE scoring, ATOM provenance tracking, Fibonacci-weighted analysis, and cross-platform translation among Claude, Grok, Gemini, and Llama strands. Typical usage starts by integrating coherence into your MCP client configuration, then launching core tools such as wave_analyze for coherence checks and bump_validate for handoff validation. Use ATOM tracking (atom_track) to maintain an auditable decision trail, and leverage gate_intention_to_execution and gate_execution_to_learning to manage structured workflow phases. The server is designed to produce a robust, multi-strand coherence map (up to 75 dimensions in this release) while protecting data integrity via the ATOM Trail when strands converge at Viviani Crossing. The setup emphasizes secure API access, modular tooling, and clear separation of concerns between analysis, provenance, and learning transitions.
How to install
Prerequisites:
- Node.js and npm installed on your system
- Basic familiarity with MCP configuration and environment variables
Install the coherence MCP package (recommended version 0.3.1):
npm install @toolate28/coherence-mcp@0.3.1
Configure your MCP client to use coherence by adding the server entry (example):
{
"mcpServers": {
"coherence": {
"command": "npx",
"args": ["-y", "@toolate28/coherence-mcp"]
}
}
}
Create and populate environment variables (examples):
# Copy example env and edit
cp .env.example .env
Required environment variables (placeholders to be replaced with real tokens):
- ATOM_AUTH_TOKEN: token for authenticated operations
- SPIRALSAFE_API_TOKEN: token required for ops tools
- WAVE_TOOLKIT_BIN: optional path to wave-toolkit CLI
Start the server (via the MCP client that reads the above config):
```bash
npm run coherence:start
If you are running directly with npx as in the Quick Install example, you can start by applying:
npx -y @toolate28/coherence-mcp
Additional notes
Tips and known considerations:
- Ensure ATOM_AUTH_TOKEN and SPIRALSAFE_API_TOKEN are kept secure and not committed to source control.
- If you rely on the Wave Toolkit, set WAVE_TOOLKIT_BIN to the local CLI path to enable wave_analyze workflows.
- The coherence server relies on the Tri-Weavon architecture; when integrating with clients, ensure Claude, Grok, Gemini strands are configured for cross-platform translation.
- When debugging, check the ATOM Trail outputs at Viviani Crossings to verify data integrity and cross-strand synchronization.
- Version pinning (0.3.1) ensures reproducible behavior; upgrade cautiously and review release notes for breaking changes.
Related MCP Servers
bytebot
Bytebot is a self-hosted AI desktop agent that automates computer tasks through natural language commands, operating within a containerized Linux desktop environment.
robloxstudio
Create agentic AI workflows in ROBLOX Studio
kindly-web-search
Kindly Web Search MCP Server: Web search + robust content retrieval for AI coding tools (Claude Code, Codex, Cursor, GitHub Copilot, Gemini, etc.) and AI agents (Claude Desktop, OpenClaw, etc.). Supports Serper, Tavily, and SearXNG.
mcpcat-typescript-sdk
MCPcat is an analytics platform for MCP server owners 🐱.
mcp-gearbox
A modern desktop application for managing Model Context Protocol (MCP) servers across multiple AI agents
mcpcat-python-sdk
MCPcat is an analytics platform for MCP server owners 🐱.