Arbor
Give your coding agent graph-native context over repositories so it can reason about structure, dependencies, and relationships instead of flat file snippets.
Overview
Arbor is a MCP server for the Build phase that exposes graph-native code intelligence so agents can reason about repository structure and relationships.
What is this MCP server?
- Graph-native model of code relationships for agent queries
- Intelligence layer positioned for repository-scale navigation
- MCP server (stdio) for Claude Code, Cursor, and compatible hosts
- Version 1.0.0 in official MCP server schema
- GitHub-hosted server definition at Anandb71/arbor
- Server schema version 1.0.0 (2025-12-11 MCP server schema)
- Source repository: github.com/Anandb71/arbor
Community signal: 122 GitHub stars.
What problem does it solve?
Agents lose accuracy on large repos when context is flat files and keyword search, not how modules actually connect.
Who is it for?
Indie builders running Claude Code or Cursor on non-trivial codebases who want relationship-aware repo context via MCP.
Skip if: Greenfield hello-world repos or teams that only need inline completions with no agent-driven architecture questions.
What do I get? / Deliverables
After you connect Arbor, your agent can use graph-oriented tooling to navigate and explain code relationships with less guesswork on multi-file work.
- Registered Arbor MCP server in your agent config
- Agent-accessible tools for graph-oriented code intelligence queries
Recommended MCP Servers
Journey fit
Arbor is installed while you are wiring agent tooling and repo-aware workflows—the Build phase is where graph-backed code intelligence pays off for implementation and refactors. Agent-tooling is the canonical shelf for MCP servers that extend how agents understand and navigate codebases, not one-off IDE plugins.
How it compares
Graph-backed code intelligence MCP server, not a linter skill or generic filesystem search plugin.
Common Questions / FAQ
Who is io.github.Anandb71/arbor for?
Solo and small-team builders using MCP-enabled agents who need richer structural context than single-file paste or ripgrep summaries.
When should I use io.github.Anandb71/arbor?
During Build when you are extending agent tooling and want graph-native queries before large refactors, dependency tracing, or cross-package changes.
How do I add io.github.Anandb71/arbor to my agent?
Add the server from the GitHub MCP definition (Anandb71/arbor) to your client’s MCP config per your host’s stdio or remote setup, then restart the agent session.