Thanks for the heart attack, Amazon!

But don't take my word for it, try it out for yourself, it is MIT licensed, and you can create new projects with it or add it to an existing project.

[1] https://github.com/dsifry/metaswarm

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!

Comments URL: https://news.ycombinator.com/item?id=46872673

Points: 1

# Comments: 0

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!

Comments URL: https://news.ycombinator.com/item?id=46864977

Points: 5

# Comments: 2

So gtg builds all of that in and deterministically determines whether or not there are any actionable comments, and thus you can block the agent from moving forward until all actionable comments are thoroughly reviewed, acted upon or acknowledged, at which point it will change state and allow the PR to be merged.

It would poll CI in loops. Miss actionable comments buried among 15 CodeRabbit suggestions. Or declare victory while threads were still unresolved.

The core problem: no deterministic way for an agent to know a PR is ready to merge.

So I built gtg (Good To Go). One command, one answer:

$ gtg 123 OK PR #123: READY CI: success (5/5 passed) Threads: 3/3 resolved

It aggregates CI status, classifies review comments (actionable vs. noise), and tracks thread resolution. Returns JSON for agents or human-readable text.

The comment classification is the interesting part — it understands CodeRabbit severity markers, Greptile patterns, Claude's blocking/approval language. "Critical: SQL injection" gets flagged; "Nice refactor!" doesn't.

MIT licensed, pure Python. I use this daily in a larger agent orchestration system — would love feedback from others building similar workflows.

Comments URL: https://news.ycombinator.com/item?id=46656759

Points: 45

# Comments: 35