drhalto

drhalto / agentmako

Public

Local-first MCP server that gives coding agents structured context packets, code/schema facts, and diagnostics - backed by a local SQLite store.

ai-coding claude-code claude-code-plugin code-intelligence codex
10
1
100% credibility
Found Apr 26, 2026 at 10 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

agentmako is a local-first codebase intelligence engine that provides a typed MCP toolset, indexing, diagnostics, and structured context packets for AI coding agents to understand code repositories before editing them.

Star Growth

See how this repo grew from 10 to 10 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is agentmako?

Agentmako is a local-first MCP server built in TypeScript that gives coding agents structured context packets, code/schema facts, and diagnostics from your repo, backed by a local SQLite store. It indexes your codebase and optional Postgres/Supabase schemas, then serves deterministic tools like reef_scout for ranked facts or context_packet for task-specific intel, all without phoning home. Run agentmako connect to attach a project, then query via CLI or dashboard—no grep rediscovery needed.

Why is it gaining traction?

It stands out by delivering agent-ready intel via the MCP protocol, letting tools like Claude Code or Cursor grab precise repo maps, routes, and diagnostics instead of hallucinating from raw files. The local-first design with CLI commands (agentmako mcp, tool call) and a web dashboard keeps everything offline and fast, while Reef facts track changes across git states. Early adopters hook it into workflows for precommit checks or schema reviews.

Who should use this?

Fullstack devs on Next.js/TS stacks with Supabase backends, needing agents to reason over routes, auth, and DB schemas without context loss. Teams building AI coding assistants or automating reviews via MCP clients like Cursor. Anyone tired of agents ignoring your repo structure.

Verdict

Worth a spin for local agent workflows—install via npm, attach a project, and test reef_scout queries. At 10 stars and 1.0% credibility, it's raw but docs and smoke tests show promise; expect rough edges until more adoption.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.