Summary

Harness 100 provides a comprehensive collection of 100 production-grade agent teams (harnesses) for Claude AI, meticulously designed to tackle complex tasks across 10 diverse professional domains. Each harness is pre-configured with specialist agents, a dedicated orchestrator skill for workflow management, and supplementary agent-extending skills, offering ready-to-use solutions in both English and Korean. This resource significantly accelerates development by delivering robust, pre-built AI agent workflows for a wide array of applications.

How It Works

The project leverages a sophisticated "Three-Layer Skill System," comprising Orchestrator, Agent-Extending, and External skills, to ensure robust team coordination and dynamic workflow management. Each harness is architected with a team of specialist agents, a central orchestrator responsible for task delegation, error handling, and cross-validation, and domain-specific skills that amplify individual agent capabilities. This layered architecture facilitates structured outputs, dependency DAGs for intelligent task ordering, and advanced error recovery strategies, enabling highly sophisticated and reliable agent team operations.

Quick Start & Requirements

To initiate usage, simply copy any desired harness directory (e.g., ko/01-youtube-production/.claude/ or en/01-youtube-production/.claude/) directly into your project's structure. The provided README does not specify any non-default prerequisites such as GPU requirements, CUDA versions, or particular Python interpreter versions.

Highlighted Details

• Features 100 production-grade agent harnesses spanning 10 distinct domains: Content Creation, Software Development & DevOps, Data & AI/ML, Business & Strategy, Education & Learning, Legal & Compliance, Health & Lifestyle, Communication & Docs, Operations & Process, and Specialized Domains.
• Each harness is equipped with 4-5 specialist agents, a core orchestrator skill, and 2-3 agent-extending skills for enhanced functionality.
• Content and structure are available in both Korean and English.
• Employs a "Three-Layer Skill System" for effective team coordination and workflow execution.
• Adheres to rigorous Quality Standards, including Agent Team Mode, deep Domain Expertise (integrating frameworks like OWASP, Bloom's Taxonomy, Porter's 5 Forces, DCF), Structured Outputs, Dependency DAGs, comprehensive Error Handling, flexible Scale Modes, defined Test Scenarios, and precise Trigger Boundaries.
• Extensive domain knowledge is embedded through frameworks such as AIDA, SOLID, Star Schema, BMC, ADDIE, IRAC, BMR, Diataxis, SIPOC, and GHG Protocol, ensuring specialized application.

Maintenance & Community

The provided README content does not contain specific information regarding notable contributors, sponsorships, community channels (such as Discord or Slack), or a public roadmap.

Licensing & Compatibility

The project is distributed under the Apache License 2.0. This license is widely recognized for its permissiveness, generally allowing for commercial use and integration into closed-source projects without significant restrictions.

Limitations & Caveats

The README does not explicitly outline specific project limitations, current development status (e.g., alpha/beta), known bugs, or unsupported platforms. A primary dependency is the Claude AI model, the operational details of which are not elaborated upon in this document. Additionally, the project's top-level "Description" field is empty.