Key takeaways
- From Dave Sifry — 18 specialized agents coordinate through a 9-phase pipeline from GitHub issue to merged PR
- Cross-model adversarial review: Claude writes code, Codex or Gemini reviews it — eliminating single-model blind spots
- Enforced quality gates: blocking state transitions mean no path from FAIL to COMMIT — coverage enforcement agents cannot bypass
- Now multi-runtime: native support for Claude Code, Gemini CLI, and Codex CLI, installed as a plugin rather than via npm
FAQ
What is Metaswarm?
Metaswarm is a multi-agent orchestration framework for Claude Code, Gemini CLI, and Codex CLI that coordinates 18 specialized agents through a structured 9-phase pipeline with TDD, adversarial review, and automated PR shepherding.
Who created Metaswarm?
Dave Sifry created Metaswarm. It builds on Steve Yegge's BEADS system and Jesse Vincent's Superpowers framework.
What makes Metaswarm different from other orchestrators?
Cross-model adversarial review (different model reviews than writes), blocking quality gates, coverage enforcement via pre-push hooks, and a Design Review Gate where 6 agents review plans in parallel.
Is Metaswarm production ready?
Metaswarm is actively maintained (last push May 2026, 308 GitHub stars as of June 2026) and is now installed as a Claude Code plugin; the npm installer is legacy. BEADS CLI is recommended but optional, and Gemini CLI and Codex CLI are supported as native runtimes.
Executive Summary
Metaswarm is Dave Sifry's multi-agent orchestration framework for Claude Code, Gemini CLI, and Codex CLI, coordinating 18 specialized agents through a 9-phase pipeline from GitHub issue to merged PR.[1] Its distinguishing features are cross-model adversarial review (Claude writes, Codex or Gemini reviews) and blocking quality gates that prevent FAIL→COMMIT transitions. Built on BEADS (now optional) for git-native issue tracking and Superpowers for foundational workflows. As of June 2026 the repo has 308 stars and 37 forks, with the latest release (v0.11.0) shipped April 1, 2026 and commits as recent as May 16, 2026.[2]
| Attribute | Value |
|---|---|
| Creator | Dave Sifry |
| Type | Open Source (MIT) |
| Package | Claude Code plugin (npm npx metaswarm init is legacy) |
| Language | Prompts + shell + BEADS (Go) |
| Runtimes | Claude Code, Gemini CLI, Codex CLI |
| Dependencies | Node.js 18+; BEADS CLI recommended |
| Adoption (Jun 2026) | 308 stars, 37 forks, ~77 npm downloads/month |
The Problem Metaswarm Solves
Claude Code is good at writing code. It is not good at building and maintaining a production codebase.[1]
Shipping production code requires research, planning, security review, design review, tests, PR creation, CI monitoring, review comment handling, and closing the loop. That's seven or eight distinct jobs. A single agent session cannot hold all of that context, and it cannot review its own work objectively.
The typical result: you become the orchestrator. You prime the agent, tell it what to build, review output, fix what it missed, create the PR, babysit CI, respond to review comments, and repeat. The agent is a fast typist, but you're still the project manager.
Product Overview
Metaswarm provides a full orchestration layer that breaks work into phases, assigns each to a specialist agent, iterates through multiple reviews, and coordinates handoffs through PR creation and shepherding.[2]
The 9-Phase Pipeline
The pipeline was consolidated from 11 phases to 9 in post-February releases (plan validation and dependency checks folded into adjacent phases).[2]
| Phase | Description |
|---|---|
| 1. Research | Researcher agent explores codebase, finds patterns and dependencies |
| 2. Plan | Architect agent creates implementation plan with tasks |
| 3. Design Review Gate | PM, Architect, Designer, Security, UX, CTO review in parallel (6 agents) |
| 4. Work Unit Decomposition | Break plan into work units with DoD items, file scopes, dependency graph |
| 5. Orchestrated Execution | Per work unit: Implement → Validate → Adversarial Review → Commit |
| 6. Final Review | Cross-unit integration check, full test suite, coverage enforcement |
| 7. PR Creation | Creates PR with structured description and test plan |
| 8. PR Shepherd | Monitors CI, handles review comments, resolves threads |
| 9. Closure & Learning | Extracts learnings back into the knowledge base |
The framework now ships 18 agents, 13 orchestration skills, and 15 slash commands (plus 12 Gemini CLI commands).[2]
The 18 Agent Personas
| Agent | Phase | Role |
|---|---|---|
| Swarm Coordinator | Meta | Assigns work to worktrees, manages parallel execution |
| Issue Orchestrator | Meta | Decomposes issues into tasks, manages phase handoffs |
| Researcher | Research | Explores codebase, discovers patterns and dependencies |
| Architect | Planning | Designs implementation plan and service structure |
| Product Manager | Review | Validates use cases, scope, and user benefit |
| Designer | Review | Reviews API/UX design and consistency |
| Security Design | Review | Threat modeling, STRIDE analysis, auth review |
| CTO | Review | TDD readiness, codebase alignment, final approval |
| Coder | Implement | TDD implementation with coverage enforcement |
| Code Reviewer | Review | Collaborative or adversarial spec compliance |
| Security Auditor | Review | Vulnerability scanning, OWASP checks |
| PR Shepherd | Delivery | CI monitoring, comment handling, thread resolution |
| Knowledge Curator | Learning | Extracts learnings, updates knowledge base |
| Test Automator | Implement | Test generation and coverage enforcement |
| Metrics | Support | Analytics and weekly reports |
| SRE | Support | Infrastructure and performance |
| Slack Coordinator | Support | Notifications and human communication |
| Customer Service | Support | User support and triage |
Key Differentiators
Cross-Model Adversarial Review
A coding agent reviewing its own output has inherent bias. Metaswarm can delegate implementation and review tasks to external AI tools — OpenAI Codex CLI and Google Gemini CLI — with one rule: the writer is always reviewed by a different model.[1]
| Pattern | Description |
|---|---|
| Cross-Model Review | If Claude writes, Codex or Gemini reviews. If Codex writes, Claude or Gemini reviews. |
| Availability-Aware Escalation | Model A (2 tries) → Model B (2 tries) → Claude (1 try) → user alert |
| Shell Adapters | Each external tool has health checks, implement, and review commands |
Blocking Quality Gates
The hardest problem is getting agents to maintain standards. Metaswarm implements blocking state transitions — there is no instruction path from FAIL to COMMIT.[1]
Three enforcement points, one config file (.coverage-thresholds.json):
| Gate | Mechanism |
|---|---|
| Pre-Push Hook | Husky git hook runs lint, typecheck, format, coverage before every push |
| CI Coverage Job | GitHub Actions workflow blocks merge on failure |
| Agent Completion Gate | Task completion checklist reads enforcement command from config |
Orchestrated Execution Loop
For complex tasks with written specs, every work unit runs through a 4-phase loop:
- Implement — Coding agent builds against spec using TDD
- Validate — Orchestrator runs tsc, eslint, vitest, coverage independently (never asks agent "did tests pass?")
- Adversarial Review — Fresh review agent checks each DoD item with file:line evidence. Binary PASS/FAIL.
- Commit — Only after adversarial PASS. If human checkpoint defined, system pauses for approval.
On failure: fix, re-validate, spawn a fresh reviewer (never the same one), retry up to 3 times before escalating.
Self-Improving Knowledge Base
Metaswarm maintains a JSONL knowledge base in your repo — patterns, gotchas, architectural decisions, anti-patterns. After every merged PR, the self-reflect workflow analyzes what happened and writes new entries.[1]
Conversation introspection watches for signals:
- You repeated yourself → candidate for new skill or command
- You disagreed → captures your preferred approach for future alignment
- You did something manually → flags as workflow automation candidate
Technical Architecture
Dependencies
| Component | Purpose |
|---|---|
| Claude Code / Gemini CLI / Codex CLI | Native AI coding runtimes (any one) |
| BEADS CLI (v0.40+, optional) | Git-native, AI-first issue tracking (bd command) — recommended, no longer required |
| Superpowers (optional) | Foundational agentic skills framework |
| GitHub CLI | PR and issue management |
| Node.js 18+ | Runtime for automation scripts |
Installation
Since v0.9.0 (February 2026), the preferred install is via the Claude Code plugin marketplace:[1]
claude plugin marketplace add dsifry/metaswarm-marketplace
claude plugin install metaswarm
The original npx metaswarm init path is now legacy — npm publishing stopped at v0.9.2 (February 27, 2026) while plugin releases continued through v0.11.0 (April 1, 2026, standalone BEADS plugin integration).[3] Existing npm installs migrate with /metaswarm:migrate. Gemini CLI users install via extension, Codex CLI via plugin marketplace.[2]
Setup detects language, framework, test runner, linter, CI — then installs and customizes everything automatically. Supports TypeScript, Python, Go, Rust, Java, Ruby, JavaScript.
GTG (Good-To-Go) Integration
Metaswarm integrates with GTG for deterministic merge readiness:[4]
- All CI checks passing
- All review comments addressed
- All discussion threads resolved
- Required approvals present
Strengths
| Strength | Description |
|---|---|
| Cross-model review | Eliminates single-model blind spots with writer ≠ reviewer rule |
| Enforced quality gates | No path from FAIL to COMMIT — agents cannot bypass coverage |
| Structured pipeline | 11 phases with clear handoffs, not ad-hoc prompting |
| Self-improving | Knowledge base grows from every PR, reduces repeated mistakes |
| Open source | Full visibility into agent definitions and workflows |
| Fresh reviewer rule | Spawns new reviewer on retry to prevent anchoring bias |
Weaknesses / Risks
| Risk | Description |
|---|---|
| BEADS learning curve | BEADS (Steve Yegge's system) is now optional, but the knowledge-base features work best with it |
| Complexity | 18 agents, 9 phases, 13 skills — significant learning curve |
| External tool setup | Cross-model review requires Codex CLI and/or Gemini CLI |
| Small community | 308 stars, a single contributor, sparse public discussion — a one-maintainer project |
| New project | First released February 2026; less battle-tested than alternatives like Gastown |
What Developers Say
Public discussion is thin — the February 2026 Show HN launch drew only 5 points and two comments, and no substantive Reddit or X threads surfaced as of June 2026. From the HN thread:[5]
"This actually looks substantial unlike most of the AI agent grifting I see on HN" — N_Lens, Hacker News[5]
"This looks amazing! Curious if you (or others) have dug into the upcoming claude swarms feature? It looks like that would let you remove the dependency on beads, as claude seems to be getting native task tracking and inter-agent messaging capabilities." — yodon, Hacker News[5]
The star trajectory (53 in February to 308 in June 2026) suggests steady organic interest despite the quiet launch.[2]
Competitive Landscape
| Tool | Approach | Agents | Cross-Model | Quality Gates |
|---|---|---|---|---|
| Metaswarm | Structured pipeline | 18 | ✅ | ✅ Blocking |
| Gastown | Parallel execution | 7 roles | — | — |
| Ralph | Simple iteration loop | 1 | — | — |
| Pythagora | IDE-integrated | 14 | — | — |
Metaswarm sits between Gastown's raw parallelism and Ralph's simplicity, emphasizing structured workflows with enforced quality over maximum agent count.
For enterprise-grade orchestration with Jira integration, signed commits, and BYOK, evaluate Tembo.
Ideal Customer
Best fit:
- Teams using Claude Code who want structured, spec-driven development
- Projects requiring enforced test coverage and quality gates
- Developers who've been burned by agents that claim "tests pass" when they don't
- Those wanting cross-model review to catch blind spots
Not ideal for:
- Users wanting simple autonomous loops (use Ralph)
- Maximum parallel agents (use Gastown)
- IDE-integrated experience (use Pythagora)
- Enterprise air-gapped deployment (use Genie)
Bottom Line
Metaswarm represents a thoughtful approach to multi-agent orchestration: instead of maximizing parallelism, it maximizes trustworthiness. The cross-model adversarial review and blocking quality gates address the core problem that agents self-certify success even when things are broken.
The 18-agent, 9-phase pipeline is complex, but that complexity maps to real development workflow stages. For teams that want structured, spec-driven development with Claude Code, Gemini CLI, or Codex CLI, Metaswarm offers a well-designed alternative to ad-hoc prompting.
| Verdict | Assessment |
|---|---|
| Recommended | Quality-conscious teams on Claude Code (or Gemini/Codex CLI) who want enforced TDD, coverage gates, and cross-model review |
| Not recommended | Users wanting simple autonomous loops, maximum parallelism, or a tool with a large community behind it — this is effectively a one-maintainer project |
| Outlook | Active through May 2026 with steady star growth (53 → 308) and a pivot to plugin distribution and multi-runtime support; release cadence slowed after April, so watch whether v0.11.0 momentum continues |
Research by Ry Walker Research • methodology
Disclosure: Author is CEO of Tembo, which offers agent orchestration as an alternative to Metaswarm.