enmanuelmag

enmanuelmag / agent-harness-kit

Public

A provider-agnostic scaffolding kit for running structured multi-agent workflows in your codebase.

ahk.cardor.dev agent agent-harness llm
29
0
100% credibility
Found May 07, 2026 at 29 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

A local toolkit that structures AI agents into teams for handling coding tasks with backlogs, logs, health gates, and a real-time dashboard.

How It Works

1
🔍 Discover the helper kit

You hear about a simple organizer that lets AI assistants team up on your coding projects with clear plans and no mess.

2
🚀 Set it up in your project

In moments, you prepare a shared workspace for your AI team complete with task lists and safety checks.

3
✏️ Tweak the rules

You set simple guidelines like health checks before work starts and folders AI can touch.

4
📋 List your tasks

You jot down jobs for the AI like 'add login' with goals to know when it's done right.

5
📊 Watch the dashboard

A live screen shows tasks moving along, what each AI did, files changed, and any hiccups.

🎉 Projects get done

Your coding work finishes smoothly with full records proving everything went perfectly.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 29 to 29 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 agent-harness-kit?

Agent-harness-kit is a TypeScript scaffolding kit for running structured multi-agent workflows in your codebase, turning free-roaming AI agents into coordinated teams with task backlogs, audit trails, and health gates. It sets up a local SQLite store for every action, file touch, and tool call, so agents like Lead, Explorer, Builder, and Reviewer stay on track across sessions. Kick off with `npx ahk init` for provider-agnostic harnesses that hook into Claude Code or OpenCode via MCP.

Why is it gaining traction?

The real-time dashboard visualizes tasks, agent timelines, file ops, and tool usage over WebSocket, making multi-agent chaos observable. Atomic task claiming via SQLite transactions avoids double-work, while customizable health checks block agents until your env is green. Switching providers preserves full history—no re-setup needed.

Who should use this?

TypeScript devs building agent-driven features in Node/Bun projects, like decomposing epics into code changes without losing state. Teams tired of chat-only agents for iterative workflows, especially backend workflows needing explorer-builder-reviewer handoffs.

Verdict

Grab it for structured agent scaffolding in TypeScript codebases—CLI shines, dashboard impresses, docs are thorough despite 29 stars and 1.0% credibility score. Early maturity means test adapters (Jira, GitHub) are coming; solid for prototypes now.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.