127 PRs to Prod this wknd with 18 AI agents: metaswarm. MIT licensed
A few weeks ago I posted about GoodToGo https://news.ycombinator.com/item?id=46656759 - a tool that gives AI agents a deterministic answer to "is this PR ready to merge?" Several people asked about the larger orchestration system I mentioned. This is that system. I got tired of being a project manager for Claude Code. It writes code fine, but shipping production code is seven or eight jobs — research, planning, design review, implementation, code review, security audit, PR creation, CI babysitting. I was doing all the coordination myself. The agent typed fast. I was still the…
What it does
In the maker’s words, at launch
A few weeks ago I posted about GoodToGo https://news.ycombinator.com/item?id=46656759 - a tool that gives AI agents a deterministic answer to "is this PR ready to merge?" Several people asked about the larger orchestration system I mentioned. This is that system. I got tired of being a project manager for Claude Code. It writes code fine, but shipping production code is seven or eight jobs — research, planning, design review, implementation, code review, security audit, PR creation, CI babysitting. I was doing all the coordination myself. The agent typed fast. I was still the bottleneck. What I really needed was an orchestrator of orchestrators - swarms of swarms of agents with deterministic quality checks. So I built metaswarm. It breaks work into phases and assigns each to a specialist swarm orchestrator. It manages handoffs and uses BEADS for deterministic gates that persist across /compact, /clear, and even across sessions. Point it at a GitHub issue or brainstorm with it (it uses Superpowers to ask clarifying questions) and it creates epics, tasks, and dependencies, then runs the full pipeline to a merged PR - including outside code review like CodeRabbit, Greptile, and Bugbot. The thing that surprised me most was the design review gate. Five agents — PM, Architect, Designer, Security, CTO — review every plan in parallel before a line of code gets written. All five must approve. Three rounds max, then it escalates to a human. I expected a rubber stamp. It catches real design problems, dependency issues, security gaps. This weekend I pointed it at my backlog. 127 PRs merged. Every one hit 100% test coverage. No human wrote code, reviewed code, or clicked merge. OK, I guided it a bit, mostly helping with plans for some of the epics. A few learnings: Agent checklists are theater. Agents skipped coverage checks, misread thresholds, or decided they didn't apply. Prompts alone weren't enough. The fix was deterministic gates — BEADS, pre-push hooks, CI jobs all on top of the agent completion check. The gates block bad code whether or not the agent cooperates. The agents are just markdown files. No custom runtime, no server, and while I built it on TypeScript, the agents are language-agnostic. You can read all of them, edit them, add your own. It self-reflects too. After every merged PR, the system extracts patterns, gotchas, and decisions into a JSONL knowledge base. Agents only load entries relevant to the files they're touching. The more it ships, the fewer mistakes it makes. It learns as it goes. metaswarm stands on two projects: https://github.com/steveyegge/beads by Steve Yegge (git-native task tracking and knowledge priming) and https://github.com/obra/superpowers by Jesse Vincent (disciplined agentic workflows — TDD, brainstorming, systematic debugging). Both were essential. Background: I founded Technorati, Linuxcare, and Warmstart; tech exec at Lyft and Reddit. I built metaswarm because I needed autonomous agents that could ship to a production codebase with the same standards I'd hold a human team to. $ cd my-project-name $ npx metaswarm init MIT licensed. IANAL. YMMV. Issues/PRs welcome!
Does the same job
all alternatives →- CACoding Agents swarming your codebaseSep 2025 · infrastructureas.ai · ▲9
I built this because I was tired of creating pull requests in 20 repositories just to change a single line of workflow job version. With Infra as AI, just mention the change. Agents work on all repos in parallel, read the docs, make a bunch of PRs and fill in the description. You can see the demo of the actual dashboard in the landing. Let me know your thoughts :) It means a lot to me!
- TMTracking Merged PRs by OpenAI's Codex and GitHub's Copilot2025 · github.com · ▲6
Figured I'd track the popular coding agents to see merge and approval rates. At first I was tracking Opened vs. Approved, but realized most people aren't approving; they just merge. I haven't found a way to reliably track the other agents (Jules, Devin, etc), since they don't have a standardized branch naming convention. Interesting to watch this over time. Open to ideas.
- LPLibretto PR agents – Automatically fix failing playwright scriptsJul 2026 · libretto.sh · ▲22
Libretto PR agents is a free TypeScript library for maintaining Playwright browser automations. Add one line of code to your existing Playwright scripts and it lets an agent automatically open GitHub PRs fixing the script when it fails. A few months ago we released Libretto, a CLI + coding-agent skill for building deterministic browser automations. The idea was that for many browser workflows, especially repetitive business workflows, you don’t need an AI agent making decisions at runtime. You want deterministic Playwright scripts that are inspectable, faster to run, and much cheaper than…
- OSOpen-source pull request review agent2024 · ▲8
Hey HN! I'm Hussam, a Senior Algorithm Engineer at CodiumAI. We launched PR Agent a little over a year ago — it’s a free and open-source tool that helps with reviewing pull requests. The response has been great and more than 70 contributors have added to the project. Today, we're launching PR Agent as a browser extension. Key features: - AI Chat in PRs: You can ask questions about specific code changes and the context is passed to the AI model - Automated Code Reviews: The /review command does an automated review, focusing on potential bugs, security issues, and best practice violations…
- WSWe scored 50k PRs with AI – what we learned about code complexityMar 2026 · gitvelocity.dev · ▲11
I'm a CTO with a ~16-person engineering team. Last year I wanted real data on what was actually shipping, not guesswork or story point theater. So we built GitVelocity. Every merged PR gets scored 0–100 by Claude across six dimensions: scope (0–20), architecture (0–20), implementation (0–20), risk (0–20), quality (0–15), perf/security (0–5). Six dimensions added up, then scaled by change size — a 10-line fix scores lower than a 500-line refactor even at the same complexity. Full formula at gitvelocity.dev/scoring-guide. After scoring 50,000+ PRs across TypeScript, Python, Rust, Go,…
- MAMulti-Agent-Coder Is #12 on Stanford's TBench. Beats Claude Code2025 · github.com · ▲5
This weekend I built a multi-agent coding system which, quite unexpectedly, beat Claude Code on Stanford's Terminal Bench! The architecture is straightforward, consisting of an orchestrator agent that deploys explorer & coder subagents to complete complex terminal based tasks, utilising an intelligent context sharing mechanism along the way which makes it all work. The repo has a lot of technical details, and all the code and prompts for you to play around with if you'd like! I had a lot of fun making this, I hope you have fun reading the README, using it yourself, or even extending it! As…
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 27d ago · cactuscompute.com
