By jyuno426
Long-running autonomous development harness with multi-agent orchestration patterns
Long-running autonomous development harness for Claude Code.
Multi-agent orchestration patterns that go the distance.
# 로컬 테스트
claude --plugin-dir /path/to/marathon
# 사용
/marathon:pge "할일 관리 앱을 만들어줘. React + Vite로."
3-agent loop: 프롬프트를 스펙으로 확장하고, 코드를 구현하고, E2E 테스트로 검증하여 품질 기준을 충족할 때까지 반복.
/marathon:pge "블로그 플랫폼. Next.js, MDX 지원, 다크모드"
흐름:
.harness/spec.md.harness/qa_report.md특징:
프로젝트에 .harness/profile.yaml을 두면 설정을 커스터마이즈할 수 있습니다:
language: ko
mode: greenfield # greenfield | extend | fix
stack:
frontend: "Next.js 15, TypeScript, Tailwind CSS"
backend: "Next.js API Routes"
evaluator:
score_threshold: 7
limits:
max_rounds: 5
없으면 디폴트로 실행됩니다.
git clone https://github.com/jyuno426/marathon.git
claude --plugin-dir ./marathon
/plugin install marathon
MIT
Based on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimnpx claudepluginhub jyuno426/marathon --plugin marathonAutonomous multi-agent development framework with spec-driven sprints and convergent iteration
Autonomous Development Orchestrator - Transform ideas into production-ready applications through multi-agent pipeline. Spec → Plan → Tasks → 100% Working App.
Plan-Execute-Review development workflow with brainstorming, executor agent, and adversarial code review loop
Structured project planning and execution through brainstorm, spec, and build phases across three execution tiers: sequential, delegated sub-agents, and full agent teams
Plan-Build-Run — Structured development workflow for Claude Code. Solves context rot through disciplined subagent delegation, structured planning, atomic execution, and goal-backward verification.
Design fluency for frontend development. 1 skill with 23 commands (/impeccable polish, /impeccable audit, /impeccable critique, etc.) and curated anti-pattern detection.