# What is the DecisionPen Context Engine (MCP)?

> Canonical: https://decisionpen.com/docs/context-engine-mcp  
> GitHub wiki: https://github.com/DecisionPen/decisionpen-mcp/blob/main/README.md

## Summary

The Context Engine is DecisionPen’s MCP server. It lets Claude, Cursor, and other SDLC agents use the Architecture Context Plane as shared truth. Agents can research that plane and propose updates. People still approve what joins the trusted Enterprise Knowledge Graph. The MCP server does not mark anything Confirmed on its own.

## Setup

- Connect Cursor or Claude: https://decisionpen.com/docs/connect-mcp-cursor-claude  
- Connect an S3 document store: https://decisionpen.com/docs/mcp-s3-document-store  

## FAQ

**Do agents auto-confirm the map?**  
No. Agents can propose updates. A person must approve them before they enter the trusted map.

**Is this a coding harness?**  
No. It is shared architecture context for agents — so they stop inventing topology and stop re-reading the same document pile on every prompt.