https://github.com/revfactory/harness
A meta-skill that designs domain-specific agent teams, defines specialized agents, and generates the skills they use.
https://github.com/revfactory/harness
claude-code claude-code-plugin harness harness-engineering
Last synced: 3 months ago
JSON representation
A meta-skill that designs domain-specific agent teams, defines specialized agents, and generates the skills they use.
- Host: GitHub
- URL: https://github.com/revfactory/harness
- Owner: revfactory
- License: apache-2.0
- Created: 2026-03-26T21:22:07.000Z (4 months ago)
- Default Branch: main
- Last Pushed: 2026-04-17T19:36:15.000Z (3 months ago)
- Last Synced: 2026-04-17T21:28:16.874Z (3 months ago)
- Topics: claude-code, claude-code-plugin, harness, harness-engineering
- Language: HTML
- Homepage: https://revfactory.github.io/harness/
- Size: 8.84 MB
- Stars: 2,585
- Watchers: 22
- Forks: 372
- Open Issues: 2
-
Metadata Files:
- Readme: README.md
- Changelog: CHANGELOG.md
- License: LICENSE
Awesome Lists containing this project
- awesome-claude-code-and-skills - revfactory/harness - square) (Development & Engineering / Specialized Agents)
- awesome-claude-code - Harness - A meta-skill that designs domain-specific agent teams, defines specialized agents, and generates the skills they use. Resources are in Korean but can produce high-quality English-language output. (Workflows & Knowledge Guides ๐ง / Teams)
- AiTreasureBox - revfactory/harness - 06-11_6791_42](https://img.shields.io/github/stars/revfactory/harness.svg)|A meta-skill that designs domain-specific agent teams, defines specialized agents, and generates the skills they use.| (Repos)
- awesome-harness-engineering - revfactory/harness - skill that generates domain-specific agent teams and the skills they use. Good example of harness-as-code, where the harness itself is produced by an agent.  (Reference Implementations / Generators & Meta-Harnesses)
- awesome-android-ai-dev-sources - ๐
- awesome-github-projects - harness - A meta-skill that designs domain-specific agent teams, defines specialized agents, and generates the skills they use. โญ8,208 `HTML` ๐ฅ (๐ค AI & Machine Learning)
- awesome-claude-code - **harness** - skill that designs domain-specific agent teams, defines specialized agents, and generates the skills they use. (๐ Claude Plugins)
- awesome-opensource-ai - Harness - Team-architecture factory for Cโlaude Code that designs domain-specific agent teams, defines specialized agents, and generates their skills. Apache 2.0 licensed.  (4. Agentic AI & Multi-Agent Systems)
- awesome-agent-harness - GitHub - 7125-f4b400?style=flat-square)](https://github.com/revfactory/harness) | claude-code, meta-factory, agent-teams | Claude Code meta-factory that generates domain-specific agent teams, skills, orchestration patterns, and validation steps from a project description. | (Catalog / Reference Harness Implementations)
- awesome-claude-code-cn - Harness - ไธไธชๅ ๆ่ฝ๏ผ็จไบ่ฎพ่ฎก็นๅฎ้ขๅ็ไปฃ็ๅข้ใๅฎไนไธไธไปฃ็ๅนถ็ๆๅฎไปฌไฝฟ็จ็ๆ่ฝใ่ตๆบๆฏ้ฉ่ฏญ็๏ผไฝๅฏไปฅไบง็้ซ่ดจ้็่ฑ่ฏญ่พๅบใ (ๅทฅไฝๆต็จไธ็ฅ่ฏๆๅ ๐ง / ๅข้)
README
# Harness โ The Team-Architecture Factory for Claude Code
**English** | [ํ๊ตญ์ด](README_KO.md) | [ๆฅๆฌ่ช](README_JA.md)
> **Harness is a team-architecture factory for Claude Code.** Say **"build a harness for this project"** (English) or **"ํ๋ค์ค ๊ตฌ์ฑํด์ค"** (ํ๊ตญ์ด) or **"ใใผใในใๆงๆใใฆ"** (ๆฅๆฌ่ช), and the plugin turns your domain description into an agent team and the skills they use โ picked from six pre-defined team-architecture patterns.
## Overview
Harness leverages Claude Code's agent team system to decompose complex tasks into coordinated teams of specialized agents. Say "build a harness for this project" and it automatically generates agent definitions (`.claude/agents/`) and skills (`.claude/skills/`) tailored to your domain.
## Category โ Where Harness Sits
Harness lives at the **L3 Meta-Factory** layer of the Claude Code ecosystem โ the layer that generates other harnesses rather than being one. Inside L3, we pick a specific sub-layer: **Team-Architecture Factory**.
| Layer | What it does | Neighbors we coexist with |
|-------|--------------|---------------------------|
| **L3 โ Meta-Factory / Team-Architecture Factory** (us) | Domain sentence โ agent team + skills, via 6 pre-defined team patterns | โ |
| L3 โ Meta-Factory / Runtime-Configuration Factory | Deterministic, repeatable runtime configurations | [coleam00/Archon](https://github.com/coleam00/Archon) |
| L3 โ Meta-Factory / Codex Runtime Port | Same concept, Codex runtime | [SaehwanPark/meta-harness](https://github.com/SaehwanPark/meta-harness) |
| L2 โ Cross-Harness Workflow | Standardize skills/rules/hooks across multiple harnesses | [affaan-m/ECC](https://github.com/affaan-m/everything-claude-code) |
> Archon generates deterministic runtime configurations. Harness generates team architectures (pipeline, fan-out/fan-in, expert pool, producer-reviewer, supervisor, hierarchical delegation) plus the skills agents use. Different sub-layers of the same L3. Pick Archon for runtime determinism, Harness for team architecture, or combine them.
## Star History
## Key Features
- **Agent Team Design** โ 6 architectural patterns: Pipeline, Fan-out/Fan-in, Expert Pool, Producer-Reviewer, Supervisor, and Hierarchical Delegation
- **Skill Generation** โ Auto-generates skills with Progressive Disclosure for efficient context management
- **Orchestration** โ Inter-agent data passing, error handling, and team coordination protocols
- **Validation** โ Trigger verification, dry-run testing, and with-skill vs without-skill comparison tests
## Harness Evolution Mechanism
The harness evolution mechanism feeds deltas (what worked / what didn't) back into the factory, so the next generation is measurably better. When a generated harness is used in a real project, the `/harness:evolve` skill captures the delta between the initial architecture and the shipped one, and feeds it back into the factory so the next generation for a similar domain starts closer to the shipped state.
```
Initial harness โโโถ Real project use โโโถ Shipped harness
โ
โผ (delta capture via /harness:evolve)
โโโโโโโโโโโโโโโโโ
โ Factory โโโโ better next-gen draft
โโโโโโโโโโโโโโโโโ
```
We call this the **Harness Evolution Mechanism** (KR: ํ๋ค์ค ์งํ ๋ฉ์ปค๋์ฆ; JA: ใใผใใน้ฒๅใกใซใใบใ ).
## Workflow
```
Phase 1: Domain Analysis
โ
Phase 2: Team Architecture Design (Agent Teams vs Subagents)
โ
Phase 3: Agent Definition Generation (.claude/agents/)
โ
Phase 4: Skill Generation (.claude/skills/)
โ
Phase 5: Integration & Orchestration
โ
Phase 6: Validation & Testing
```
## Installation
### Via Marketplace
#### Add the marketplace
```shell
/plugin marketplace add revfactory/harness
```
#### Install the plugin
```shell
/plugin install harness@harness
```
### Direct Installation as Global Skill
```shell
# Copy the skills directory to ~/.claude/skills/harness/
cp -r skills/harness ~/.claude/skills/harness
```
## Plugin Structure
```
harness/
โโโ .claude-plugin/
โ โโโ plugin.json # Plugin manifest
โโโ skills/
โ โโโ harness/
โ โโโ SKILL.md # Main skill definition (6-Phase workflow)
โ โโโ references/
โ โโโ agent-design-patterns.md # 6 architectural patterns
โ โโโ orchestrator-template.md # Team/subagent orchestrator templates
โ โโโ team-examples.md # 5 real-world team configurations
โ โโโ skill-writing-guide.md # Skill authoring guide
โ โโโ skill-testing-guide.md # Testing & evaluation methodology
โ โโโ qa-agent-guide.md # QA agent integration guide
โโโ README.md
```
## Usage
Trigger in Claude Code with prompts like:
```
Build a harness for this project
Design an agent team for this domain
Set up a harness
```
### Execution Modes
| Mode | Description | Recommended For |
|------|-------------|-----------------|
| **Agent Teams** (default) | TeamCreate + SendMessage + TaskCreate | 2+ agents requiring collaboration |
| **Subagents** | Direct Agent tool invocation | One-off tasks, no inter-agent communication needed |
### Architecture Patterns
| Pattern | Description |
|---------|-------------|
| Pipeline | Sequential dependent tasks |
| Fan-out/Fan-in | Parallel independent tasks |
| Expert Pool | Context-dependent selective invocation |
| Producer-Reviewer | Generation followed by quality review |
| Supervisor | Central agent with dynamic task distribution |
| Hierarchical Delegation | Top-down recursive delegation |
## Output
Files generated by Harness:
```
your-project/
โโโ .claude/
โ โโโ agents/ # Agent definition files
โ โ โโโ analyst.md
โ โ โโโ builder.md
โ โ โโโ qa.md
โ โโโ skills/ # Skill files
โ โโโ analyze/
โ โ โโโ SKILL.md
โ โโโ build/
โ โโโ SKILL.md
โ โโโ references/
```
## Use Cases โ Try These Prompts
Copy any prompt below into Claude Code after installing Harness:
**Deep Research**
```
Build a harness for deep research. I need an agent team that can investigate
any topic from multiple angles โ web search, academic sources, community
sentiment โ then cross-validate findings and produce a comprehensive report.
```
**Website Development**
```
Build a harness for full-stack website development. The team should handle
design, frontend (React/Next.js), backend (API), and QA testing in a
coordinated pipeline from wireframe to deployment.
```
**Webtoon / Comic Production**
```
Build a harness for webtoon episode production. I need agents for story
writing, character design prompts, panel layout planning, and dialogue
editing. They should review each other's work for style consistency.
```
**YouTube Content Planning**
```
Build a harness for YouTube content creation. The team should research
trending topics, write scripts, optimize titles/tags for SEO, and plan
thumbnail concepts โ all coordinated by a supervisor agent.
```
**Code Review & Refactoring**
```
Build a harness for comprehensive code review. I want parallel agents
checking architecture, security vulnerabilities, performance bottlenecks,
and code style โ then merging all findings into a single report.
```
**Technical Documentation**
```
Build a harness that generates API documentation from this codebase.
Agents should analyze endpoints, write descriptions, generate usage
examples, and review for completeness.
```
**Data Pipeline Design**
```
Build a harness for designing data pipelines. I need agents for schema
design, ETL logic, data validation rules, and monitoring setup that
delegate sub-tasks hierarchically.
```
**Marketing Campaign**
```
Build a harness for marketing campaign creation. The team should research
the target market, write ad copy, design visual concepts, and set up
A/B test plans with iterative quality review.
```
## Coexistence โ Harness and Neighbors
Harness is not alone in the Claude Code / agent-framework ecosystem. The following repos live in adjacent layers; each is described in a parallel "X is โฆ, Harness is โฆ" form so you can pick the one that fits your need or combine several.
| Repo | Their position | Relationship to Harness |
|------|----------------|-------------------------|
| [coleam00/Archon](https://github.com/coleam00/Archon) | "harness builder" โ deterministic, repeatable runtime configurations | **Same L3, neighbor sub-layer.** Archon is a Runtime-Configuration Factory, Harness is a Team-Architecture Factory. Pick Archon for runtime determinism, Harness for team architecture, or combine them. |
| [SaehwanPark/meta-harness](https://github.com/SaehwanPark/meta-harness) | Codex port of the same concept | **Same L3, different runtime.** Use Harness on Claude Code, meta-harness on Codex. |
| [affaan-m/ECC](https://github.com/affaan-m/everything-claude-code) | "Agent harness performance & workflow layer" (sits on top of existing harnesses) | **Different layer.** ECC is a standardization layer across harnesses; Harness is a factory that generates harnesses. Serial combination possible. |
| [wshobson/agents](https://github.com/wshobson/agents) | Subagent / skill catalog (182 agents, 149 skills) | **Factory โ parts supply.** wshobson is a catalog to shop from; Harness designs the team. Absorb wshobson entries as parts inside a Harness-generated team. |
| [LangGraph](https://langchain-ai.github.io/langgraph/) | State-graph orchestration, LLM-agnostic | **Different track.** LangGraph is for long-running, state-recoverable orchestration; Harness is for fast Claude-Code-native team design. |
## Built with Harness
### Harness 100
**[revfactory/harness-100](https://github.com/revfactory/harness-100)** โ 100 production-ready agent team harnesses across 10 domains, available in both English and Korean (200 packages total). Each harness ships with 4-5 specialist agents, an orchestrator skill, and domain-specific skills โ all generated by this plugin. 1,808 markdown files covering content creation, software development, data/AI, business strategy, education, legal, health, and more.
### Research: A/B Testing Harness Effectiveness
**[revfactory/claude-code-harness](https://github.com/revfactory/claude-code-harness)** โ A controlled experiment across 15 software engineering tasks measuring the impact of structured pre-configuration on LLM code agent output quality.
| Metric | Without Harness | With Harness | Improvement |
|--------|:-:|:-:|:-:|
| Average Quality Score | 49.5 | 79.3 | **+60%** |
| Win Rate | โ | โ | **100%** (15/15) |
| Output Variance | โ | โ | **-32%** |
Key finding: effectiveness scales with task complexity โ the harder the task, the greater the improvement (+23.8 Basic, +29.6 Advanced, +36.2 Expert).
**Exact phrasing to use everywhere:** +60% avg quality (49.5 โ 79.3), 15/15 win-rate, โ32% variance (n=15, author-measured A/B, third-party replications pending).
> Full paper: *Hwang, M. (2026). Harness: Structured Pre-Configuration for Enhancing LLM Code Agent Output Quality.*
## Requirements
- [Agent Teams enabled](https://code.claude.com/docs/en/agent-teams): `CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1`
## FAQ
Q1. Isn't "+60%" oversold?
**A.** The +60% figure comes from an **author-measured A/B (n=15, 15 tasks, measured on the sister repo `claude-code-harness`)**. Every citation in this repo pairs the number with the disclosure "n=15, author-measured, third-party replications pending" in the same sentence. For adoption decisions, we recommend running a 2โ4 week internal pilot and measuring your own numbers.
**Evidence:**
- Author A/B: [revfactory/claude-code-harness](https://github.com/revfactory/claude-code-harness)
- Paper: *Hwang, M. (2026). Harness: Structured Pre-Configuration for Enhancing LLM Code Agent Output Quality*
Q2. Why "harness factory" and not "harness builder"? Isn't this competing with Archon?
**A.** Archon generates deterministic runtime configurations โ it's a **Runtime-Configuration Factory**. Harness generates agent team architectures (team structure, message protocols, review gates) โ it's a **Team-Architecture Factory**. They are **neighbor sub-layers of the same L3 Meta-Factory** and serve different needs. Pick Archon for runtime determinism, Harness for team-architecture patterns, or combine them (design architecture with Harness โ deploy runtime with Archon).
**Evidence:**
- Archon self-definition: [clawfit docs/reference-levels.md](https://github.com/hongsw/clawfit/blob/main/docs/reference-levels.md)
- Sub-layer declaration: see the **Category โ Where Harness Sits** section above
- Archon repo: [github.com/coleam00/Archon](https://github.com/coleam00/Archon)
Q3. Isn't "Claude Code only" too narrow? What about Gemini/Codex?
**A.** Currently the official runtime is Claude Code only. A Codex port of the same concept โ [SaehwanPark/meta-harness](https://github.com/SaehwanPark/meta-harness) โ is already public, so Codex teams can start there. Harness chose "Claude-Code-native, deep" over "multi-runtime, shallow"; cross-runtime collaboration with sibling repos (meta-harness, harness-init, OpenRig) is on the roadmap.
**Evidence:**
- Codex port: [github.com/SaehwanPark/meta-harness](https://github.com/SaehwanPark/meta-harness)
- Cross-runtime scaffolder: [github.com/Gizele1/harness-init](https://github.com/Gizele1/harness-init)
## License
Apache 2.0