# Harness 100 Production-grade agent team harness collection for [Claude Code](https://docs.anthropic.com/en/docs/claude-code). 100 ready-to-use harnesses across 10 domains — each with 4-5 specialist agents, an orchestrator skill, and 2-3 agent-extending skills. Available in **Korean** and **English**. > **[English (en/)](en/)** | **[Korean (ko/)](ko/)** ## At a Glance | | ko/ | en/ | Total | |---|-----|-----|-------| | Harnesses | 100 | 100 | 200 | | Agent definitions | 489 | 489 | 978 | | Skills | 315 | 315 | 630 | | Total .md files | 904 | 904 | **1,808** | ## Quick Start ```bash # Copy any harness into your project cp -r ko/01-youtube-production/.claude/ /path/to/my-project/.claude/ # Or use the English version cp -r en/01-youtube-production/.claude/ /path/to/my-project/.claude/ ``` ## Categories | # | Category | Harnesses | Highlights | |---|----------|-----------|------------| | 1 | Content Creation | 01-15 | YouTube, podcast, game narrative, comics, translation | | 2 | Software Dev & DevOps | 16-30 | Full-stack, API, CI/CD, security audit, IaC | | 3 | Data & AI/ML | 31-42 | ML experiments, NLP, RAG/LLM apps, design systems | | 4 | Business & Strategy | 43-55 | Startup, market research, pricing, financial modeling | | 5 | Education & Learning | 56-65 | Language tutor, exam prep, debate simulator, ADR | | 6 | Legal & Compliance | 66-72 | Contracts, patents, GDPR/PIPA, regulatory filing | | 7 | Health & Lifestyle | 73-80 | Meal planning, fitness, tax, travel, wedding | | 8 | Communication & Docs | 81-88 | Technical writing, SOP, proposals, crisis comms | | 9 | Operations & Process | 89-95 | Hiring, onboarding, audit, procurement | | 10 | Specialized Domains | 96-100 | Real estate, e-commerce, ESG, IP portfolio | ## Harness Architecture ``` {NN}-{harness-name}/ └── .claude/ ├── CLAUDE.md # Project overview ├── agents/ │ ├── {specialist-1}.md # Domain expert agent │ ├── {specialist-2}.md │ ├── {specialist-3}.md │ ├── {specialist-4}.md │ └── {reviewer/qa}.md # Cross-validation agent └── skills/ ├── {orchestrator}/ │ └── skill.md # Team orchestration ├── {domain-skill-1}/ │ └── skill.md # Agent-extending skill └── {domain-skill-2}/ └── skill.md # Agent-extending skill ``` ### Three-Layer Skill System | Layer | Purpose | Example | |-------|---------|---------| | **Orchestrator** | Team coordination, workflow, error handling | `youtube-production/skill.md` | | **Agent-Extending** | Domain knowledge that amplifies agent expertise | `hook-writing/skill.md`, `thumbnail-psychology/skill.md` | | **External** | Existing tools (Gemini image gen, web search) | `gemini-3-pro-imagegen` | ## Quality Standards Every harness includes: - **Agent Team Mode** — SendMessage direct communication, cross-validation - **Domain Expertise** — Real frameworks (OWASP, Bloom's Taxonomy, Porter's 5 Forces, DCF, etc.) - **Structured Outputs** — Domain-specific templates per agent - **Dependency DAG** — Task ordering with parallel execution - **Error Handling** — Retry, skip, fallback strategies - **Scale Modes** — Full pipeline / reduced / single-agent - **Test Scenarios** — Normal / existing-file / error (3 types) - **Trigger Boundaries** — Should-trigger + NOT-trigger defined ## Domain Expertise | Domain | Embedded Frameworks | |--------|-------------------| | Content | AIDA, Pattern Interrupt, CURVE formula, Platform Specs | | Development | SOLID, DDD, OWASP Top 10, Test Pyramid, DORA Metrics, CWE Top 25 | | Data | Star/Snowflake Schema, Great Expectations, SHAP/LIME, Feature Engineering | | Business | BMC, TAM/SAM/SOM, Porter's 5 Forces, RICE, Van Westendorp PSM | | Education | Bloom's Taxonomy, ADDIE, CEFR, SM-2 Spaced Repetition, Toulmin Model | | Legal | IRAC, MQM, GDPR/PIPA, IPC/CPC, Claim Drafting Patterns | | Lifestyle | BMR/TDEE, ACSM Guidelines, Compound Interest, Route Optimization | | Documents | Diataxis, PREP, STAR, MADR, SemVer, Mermaid Patterns | | Operations | SIPOC/RACI, 4C Framework, SMART, NPS/CSAT, BARS Assessment | | Specialized | GHG Protocol, Cap Rate/IRR, IMRaD, Georgia-Pacific, Double Materiality | ## License Apache License 2.0 — See [LICENSE](LICENSE)