Gaffer — AI skill for Claude Code
Self-hosted AI coding factory — sandboxed agents deliver tickets to merged code, gated by a human in a dashboard.
How to install Gaffer
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open tmj-90/gaffer and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Gaffer does
Self-hosted AI coding factory — sandboxed agents deliver tickets to merged code, gated by a human in a dashboard. Local-first, cost-transparent, human-in-the-loop.
Alternatives in AI
- Omnara — by Ishaan Sehgal - A command center for AI agents that syncs Claude Code sessions across terminal, web, and mo 2.6k ★
- Agentrq — AgentRQ: Human-in-loop realtime conversational task manager for AI Agents 1.1k ★
- Openclaw Dashboard — 🔐 Secure, real-time monitoring dashboard for OpenClaw AI agents 654 ★
README
Gaffer
[](https://github.com/tmj-90/gaffer/actions/workflows/ci.yml) [](https://opensource.org/licenses/Apache-2.0)
**A local-first, self-hosted AI coding factory that works your own backlog — autonomous coding agents you supervise, on your machine, under a human merge gate.**
I built Gaffer to chip away at my own side-project backlog while I'm doing something else: the boring, never-quite-urgent tickets I'd otherwise never get to. It turns a one-line idea into dependency-ordered tickets, works each one through **plan → implement → test → review**, and delivers a git branch or PR with evidence — but **nothing merges until I approve the diff** (the agent *structurally cannot* ship its own work). Vague or blocked tickets park for a human rather than being forced through.
It's a personal hacking tool, not a claim that it replaces engineers or ships production code unsupervised — it's **run-at-your-own-risk alpha**. Think of it as a tireless junior for the stuff you'd never get around to: local-first (your machine, your repos, your keys), supervised by default, hands-off autonomy strictly opt-in.
**Alpha software.** Gaffer runs local coding agents with **shell access** inside repos you control. Use it only on **trusted repositories**, keep the **human review gate on**, and treat the host like a machine running an AI agent with developer-level access. Not prod
Related Skills
AI Control Plane Public
Self-hosted multi-agent AI orchestration with governance built in: one router in, cost tiers and isolated lane
ProductsFactory
Autonomous AI dev factory: sandboxed Claude Code agents design, build, review, and ship features across repos
Salient Core
Guardrail and permission layer for AI agent harnesses. A default-deny control kernel that gates every tool cal
Agentic Review Gate
Governed, auditable AI review platform: the review and approval work a team does by hand, as a webhook-trigger
AI Software Factory
Self-hosted, open-source AI Software Factory: a Governance Controller that orchestrates heterogeneous coding a
Team1 Factory
Open-source AI software factory: GitHub issues in, merged PRs out.
Related Agents
Tsk - AI Agent Task Manager And Sandbox
by dtormoen - A Rust CLI tool that lets you delegate development tasks to AI agents running in sandboxed Docke
AI Finance Engineer
AI Finance & Cost Optimization specialist who manages AI infrastructure costs, builds financial models for AI
Img2threejs Agent
Procedural image-to-Three.js reconstruction specialist from img2threejs/img2threejs. Rebuilds a reference obje