AgentMachinist
A local-first CLI that turns a GitHub issue into a reviewed, tested pull request through a spec-approval pipeline.
🔗 Visit AgentMachinist
Description
Handing a GitHub issue to an AI coding agent and hoping for a good pull request back is a gamble — the agent might misunderstand the ask, skip tests, or produce something nobody reviewed properly. AgentMachinist is built to make that process disciplined: it takes a GitHub issue, walks it through specification, implementation, and testing, and gates the resulting pull request behind an approval step tied to a specific commit (SHA-bound), so what got approved is exactly what ships.
It's local-first — running on your own machine rather than a cloud service — and works across several AI coding harnesses (Claude, OpenCode, Pi, Codex) instead of locking you into one vendor. It ships as a CLI (the machinist command) with task queue controls (pause, resume, defer), a repository registry for managing multiple projects, and diagnostic tooling. It's free, open source, and installable via PyPI.
💬 Our review
The short version: a disciplined, solo-developer-focused alternative to just trusting an AI agent to open a PR — the SHA-bound approval step is the standout idea, forcing a real checkpoint between 'agent proposed a spec' and 'agent starts writing code.'
It sits alongside tools like GitHub Copilot Workspace, Devin, and Cursor's background agents, all of which automate issue-to-PR to varying degrees, but most of those are hosted/cloud products with subscription pricing. AgentMachinist's local-first, multi-harness, free/open-source approach is the differentiator if you want that workflow without handing your repo to a third-party service. The trade-off is you're managing the orchestration yourself via CLI rather than getting a polished hosted dashboard.
💰 Pricing
📊 Global score
🤖 AI-enriched data
Open source, disponible via PyPI, sans coût.
Pros
Validation liée à un SHA précis, vrai point de contrôle avant le code
Local-first, compatible plusieurs harnais IA (Claude, OpenCode, Pi, Codex)
Gratuit, open source, installable via PyPI
Contrôles de file de tâches (pause/reprise/report)
Cons
CLI uniquement, pas de tableau de bord hébergé
Pensé pour un développeur solo plus que pour de grosses équipes
Demande de gérer soi-même la configuration du harnais IA