owainlewis

owainlewis / agent-worker

Public

Secure and scalable way to delegate work to AI agents from anywhere. Build a team of agents that work for you.

17
8
100% credibility
Found Mar 21, 2026 at 17 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

Agent-worker is a tool that automatically monitors an issue tracker for ready tasks, uses AI agents to complete coding work in a repository, runs custom setup and cleanup steps, and updates task statuses with results.

How It Works

1
🔍 Discover the helper

You hear about agent-worker, a smart tool that watches your task list and finishes coding jobs automatically so you don't have to babysit.

2
📥 Bring it home

Download the ready-to-use program and set it up on your computer.

3
🔗 Connect your workspace

Link it to your task tracker and code project folder, sharing access so it can see tasks and make changes.

4
⚙️ Personalize it

Decide how often it checks for work, pick your favorite AI thinker, and add simple steps like saving changes or sharing updates.

5
🚀 Turn it on

Start the worker with a simple command, and it quietly begins monitoring everything on its own schedule.

6
🤖 Let it handle tasks

It spots ready tasks, claims them, uses AI to do the coding in a safe space, runs your steps, and reports back.

✅ Work done, freedom gained

Tasks move to completed, with updates and new code ready for review, freeing you to focus on what matters.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 17 to 17 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-worker?

Agent-worker is a TypeScript tool built with Bun that automates dev workflows by polling your Linear project for "ready" tickets, claiming them, and delegating the work to AI agents like Claude Code or OpenAI Codex. It runs the agent in an isolated git worktree, executes configurable pre- and post-hooks (like testing, committing, pushing, and creating PRs via GitHub CLI), then updates the ticket status and adds comments—all without you intervening. This solves the babysitting problem of AI agents by making them fully autonomous in a secure, polling-based setup.

Why is it gaining traction?

Its polling model keeps things secure and scalable—no exposed webhooks or inbound ports, just outbound checks every 60 seconds by default, letting you run multiple agent workers in parallel against the same repo without conflicts. Hooks provide deterministic guardrails around flaky AI output, and YAML config with variable interpolation makes setup straightforward for custom pipelines. Agent-agnostic design means adding new executors is simple, standing out from webhook-heavy orchestrators.

Who should use this?

Indie hackers or small teams on Linear who assign simple coding tasks like "add login form" or "fix bug in API route" and want AI to handle implementation-to-PR autonomously. Ideal for backend devs using Bun/TypeScript stacks with CLAUDE.md for project conventions, or anyone tired of manual ticket-to-agent copy-pasting.

Verdict

Early days with 17 stars and 1.0% credibility score—maturity is low, but solid README and example config make it easy to spin up. Try it on a side project if you match the stack; skip for production until more battle-testing.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.