I.
Wiki article
Reading · 4 minlibrary/metaswarm
Metaswarm Methodology (Library) reference
Metaswarm is an autonomous multi-agent orchestration framework that manages the complete lifecycle from GitHub issue to merged pull request. It coordinates 12 specialized agents through a rigorous 7-phase workflow with quality gates as blocking state transitions, adversarial reviews with fresh reviewers, and knowledge persistence across sessions.
Continue reading
Nearby pages in the same section.
Aerospace Engineering Specialization (Library)AI Agents and Conversational AI Specialization (Library)Algorithms and Optimization Specialization (Library)Arts and Culture Specialization (Library)ATDD/TDD Methodology (Library)AutoMaker (Library)Automotive Engineering Specialization (Library)Backend Development (Library)BDD/Specification by Example (Library)Bioinformatics and Genomics Specialization (Library)Biomedical Engineering Specialization (Library)BMAD Method (Library)Business Analysis and Consulting (Library)Business Strategy and Operations (Library)Business Strategy Specialization (Library)CC10X Methodology (Library)CCPM - Claude Code PM Methodology (Library)Chemical Engineering Specialization (Library)Civil Engineering Specialization (Library)ClaudeKit Methodology (Library)Cleanroom Software Engineering (Library)CLI and MCP Development Specialization (Library)Code Migration and Modernization Specialization (Library)COG Second Brain (Library)Common Utilities (Library)Computer Science Specialization (Library)Cryptography and Blockchain Development Specialization (Library)Customer Experience and Support Specialization (Library)Data Engineering, Analytics, and BI Specialization (Library)Data Science and Machine Learning Specialization (Library)Intelligence, Decision Support and Decision Making (Library)Desktop Product Development Specialization (Library)DevOps, SRE, and Platform Engineering Specialization (Library)Digital Marketing and Content Strategy Specialization (Library)Domain-Driven Design (DDD) Methodology (Library)Double Diamond Methodology (Library)Education and Learning Specialization (Library)Electrical Engineering Specialization (Library)Embedded Systems Engineering Specialization (Library)Entrepreneurship and Startup Processes (Library)Environmental Engineering Specialization (Library)Event Storming (Library)Everything Claude Code Methodology (Library)Example Mapping Methodology (Library)Extreme Programming (XP) (Library)Feature-Driven Development (FDD) (Library)Finance, Accounting, and Economics Specialization (Library)FPGA Programming and Hardware Description Specialization (Library)Game Product Development Specialization (Library)Gas Town Methodology (Library)GPU Programming and Parallel Computing (Library)GSD-Adapted Workflows for Babysitter SDK (Library)Healthcare and Medical Management Specialization (Library)Human Resources and People Operations Specialization (Library)Humanities and Anthropology Specialization (Library)Hypothesis-Driven Development (Library)Impact Mapping Methodology (Library)Industrial Engineering Specialization (Library)Jobs to Be Done (JTBD) Methodology (Library)Kanban (Library)Knowledge Management (Library)Legal and Compliance Specialization (Library)Logistics and Operations Specialization (Library)Maestro App Factory (Library)Marketing and Brand Management Specialization (Library)Materials Science Specialization (Library)Mathematics Specialization (Library)Mechanical Engineering Specialization (Library)Meta Specialization - Process, Skill, and Agent Creation (Library)Mobile Product Development Specialization (Library)Nanotechnology Specialization (Library)Network Programming and Protocols Specialization (Library)Enhanced Ontology-Driven Development (ODD) Methodology (Library)Operations Management Specialization (Library)Performance Optimization and Profiling Specialization (Library)Philosophy and Theology Specialization (Library)Physics Specialization (Library)Pilot Shell Methodology for Babysitter SDK (Library)Planning with Files (Library)Product Management and Product Strategy Specialization (Library)Production contract (Library)Programming Languages and Compilers Development Specialization (Library)Project Management and Leadership Specialization (Library)Public Relations and Communications Specialization (Library)QA, Testing, and Test Automation (Library)Quantum Computing Specialization (Library)Research Specialization (Library)Robotics and Simulation Engineering Specialization (Library)RPIKit Methodology (Library)Ruflo Methodology (Library)RUP (Rational Unified Process) (Library)Sales and Business Development Specialization (Library)Scientific Discovery and Problem Solving Specialization (Library)Scrum (Library)SDK, Platform, and Systems Development (Library)Security, Compliance, and Risk Management Specialization (Library)Security Research and Vulnerability Analysis Specialization (Library)Shape Up (Library)Social Sciences Specialization (Library)Software Architecture and Design Patterns Specialization (Library)Spec Kit Methodology (Library)Spiral Model (Library)Superpowers Extended Methodology (Library)Supply Chain Management Specialization (Library)Technical Documentation Specialization (Library)Travel (Curated-Dataset + SQL-Tool Pattern) (Library)UX/UI Design and User Experience Specialization (Library)V-Model Methodology (Library)Venture Capital and Investment Due Diligence Specialization (Library)Waterfall Methodology (Library)Web Product Development Specialization (Library)
Documented graph nodes
Records linked directly from this page’s Page node.
Metaswarm Methodology
**Source**: dsifry/metaswarm by David Sifry **Category**: Autonomous Multi-Agent Orchestration / Issue-to-PR Lifecycle **License**: See upstream repository
Overview
Metaswarm is an autonomous multi-agent orchestration framework that manages the complete lifecycle from GitHub issue to merged pull request. It coordinates 12 specialized agents through a rigorous 7-phase workflow with quality gates as blocking state transitions, adversarial reviews with fresh reviewers, and knowledge persistence across sessions.
Core Principles
- **Trust Nothing, Verify Everything, Review Adversarially** - Quality gates are blocking, never advisory
- **TDD Mandatory** - Write tests first, watch them fail, then implement (100% coverage targets)
- **Fresh Reviewer Rule** - On re-review after FAIL, spawn new reviewer with no memory (prevents anchoring bias)
- **Independent Validation** - Orchestrator runs quality gates directly, never trusts subagent self-reports
- **Human Checkpoints** - Planned pauses at critical boundaries (schema changes, security code, new patterns)
- **Knowledge Persistence** - Extract learnings while context is fresh, persist for cross-session continuity
Process Files
| Process | File | Description | Task Count |
|---|---|---|---|
| Issue Orchestrator | metaswarm-orchestrator.js | Full 7-phase lifecycle: research to PR | 12 |
| Design Review Gate | metaswarm-design-review.js | 6 parallel specialist reviews, unanimous approval | 7 |
| Execution Loop | metaswarm-execution-loop.js | 4-phase cycle: Implement -> Validate -> Review -> Commit | 4 |
| Swarm Coordinator | metaswarm-swarm-coordinator.js | Multi-issue parallel management across worktrees | 6 |
| Knowledge Cycle | metaswarm-knowledge-cycle.js | Context priming and self-reflection | 5 |
| PR Shepherd | metaswarm-pr-shepherd.js | PR lifecycle through merge | 3 |
Skills Catalog
| Skill | Directory | Description |
|---|---|---|
| orchestrated-execution | skills/orchestrated-execution/ | 4-phase execution loop with quality gates |
| design-review-gate | skills/design-review-gate/ | 6-agent parallel design review |
| plan-review-gate | skills/plan-review-gate/ | 3 adversarial plan reviewers |
| adversarial-review | skills/adversarial-review/ | Fresh reviewer with binary PASS/FAIL |
| knowledge-curation | skills/knowledge-curation/ | Context priming and self-reflection |
| work-unit-decomposition | skills/work-unit-decomposition/ | DoD items, file scope, dependencies |
| pr-shepherding | skills/pr-shepherding/ | PR lifecycle management through merge |
| external-tool-coordination | skills/external-tool-coordination/ | Cross-model AI tool integration |
Agents Catalog
| Agent | Directory | Role |
|---|---|---|
| issue-orchestrator | agents/issue-orchestrator/ | Master coordinator per issue |
| researcher | agents/researcher/ | Codebase exploration and analysis |
| architect | agents/architect/ | Implementation planning and decomposition |
| product-manager | agents/product-manager/ | Use case and scope validation |
| designer | agents/designer/ | UX/API design review |
| security-design | agents/security-design/ | Threat modeling and OWASP analysis |
| cto | agents/cto/ | TDD readiness and codebase alignment |
| coder | agents/coder/ | TDD implementation specialist |
| code-reviewer | agents/code-reviewer/ | Fresh adversarial reviewer |
| security-auditor | agents/security-auditor/ | Implementation security review |
| pr-shepherd | agents/pr-shepherd/ | PR lifecycle management |
| swarm-coordinator | agents/swarm-coordinator/ | Multi-issue parallel orchestration |
Workflow Lifecycle
Issue -> Research -> Plan -> Plan Review Gate (3 adversarial) -> Preflight -> Design Review Gate (6 parallel, unanimous) -> Work Unit Decomposition -> Execution Loop (Implement -> Validate -> Adversarial Review -> Commit) x N -> Final Review -> Self-Reflect -> PR -> Shepherd -> MergeQuality Gates (Blocking)
1. **Plan Review Gate** - 3 adversarial reviewers (Feasibility, Completeness, Scope) 2. **Design Review Gate** - 6 parallel specialists (ALL must approve, max 3 iterations) 3. **Quality Gate Validation** - Independent tsc, eslint, vitest (blocking, never advisory) 4. **Adversarial Code Review** - Fresh reviewer, binary PASS/FAIL (max 3 attempts -> escalate) 5. **Coverage Gate** - 100% target across lines/branches/functions/statements
Anti-Patterns (Enforced)
- Self-certifying (trusting subagent claims)
- Combining execution phases into single steps
- Reusing reviewers after FAIL verdict
- Passing previous review findings to new reviewers
- Treating quality gate failures as advisory
- Proceeding past human checkpoints without explicit approval
- Using --no-verify on commits or force-pushing
Philosophy
- **Adversarial over collaborative** review for verified compliance
- **Independent validation** over trusted reporting
- **Blocking gates** over advisory recommendations
- **Fresh reviewers** over experienced reviewers (prevent bias)
- **Knowledge extraction** before context is lost
- **Human escalation** after bounded retry (3 attempts)