Agentcontext
Maintain one source of truth and emit AGENTS.md, CLAUDE.md, Cursor, Cline, and Windsurf rule files without manual copy-paste drift.
Overview
agentcontext is a MCP server for the Build phase that normalizes AGENTS.md, CLAUDE.md, and editor-specific rule files from one shared source.
What is this MCP server?
- Normalizes AGENTS.md, CLAUDE.md, .cursor/rules, .clinerules, and .windsurf/rules
- Single-source workflow to stop conflicting instructions across editors
- npm package @mbeato/agentcontext-mcp over stdio MCP
- Targets multi-agent solo stacks (Claude Code + Cursor + Windsurf)
- Version 0.1.2 with GitHub source at mbeato/agentcontext-mcp
- Supports 5 target artifact families: AGENTS.md, CLAUDE.md, .cursor/rules, .clinerules, .windsurf/rules
- npm package @mbeato/agentcontext-mcp version 0.1.2
- stdio MCP transport
What problem does it solve?
Solo builders juggling Claude Code, Cursor, and Windsurf end up with conflicting AGENTS and rules files that make every agent session behave differently.
Who is it for?
Indie devs running multiple AI editors on one codebase who want synchronized AGENTS.md and rules without manual duplication.
Skip if: Builders who only use a single agent with one static CLAUDE.md and never touch Cursor, Cline, or Windsurf configs.
What do I get? / Deliverables
After install, you can regenerate aligned agent instruction files from one source so every client follows the same constraints and workflows.
- Aligned AGENTS.md, CLAUDE.md, and editor rule trees generated from one source
- Reduced drift between Claude Code, Cursor, Cline, and Windsurf sessions
- Repeatable normalization step in onboarding or CI
Recommended MCP Servers
Journey fit
Canonical shelf is Build agent-tooling because you wire agent behavior while setting up the repo, but normalized rules pay off across Ship review and Operate iteration too. Agent-tooling fits best: the server’s job is synchronizing multi-client agent instruction files from a single normalized definition.
How it compares
MCP normalization layer for agent instruction files, not a database connector or paid SEO API suite.
Common Questions / FAQ
Who is agentcontext for?
Solo builders and small teams standardizing AI behavior across Claude Code, Cursor, Cline, and Windsurf from one canonical agent context.
When should I use agentcontext?
Use it when you add a new agent client, refactor coding policies, or notice drift between AGENTS.md and editor-specific rules folders.
How do I add agentcontext to my agent?
Install @mbeato/agentcontext-mcp, add the stdio MCP server entry in your client config, then run normalization tools against your source definitions per the repo README.