#8080th of 85 in Skills, plugins and rules
harness
Claude Code plugin that generates agent teams and skills from a project description
- Stars
- 9.1k
- License
- Apache-2.0
- Last commit
- Sep 2026
Missed the latest GitHub check; these facts may be older.
Overview
Harness is a Claude Code plugin that takes a domain description and writes agent definitions into .claude/agents/, skills into .claude/skills/, and a minimal CLAUDE.md pointer. It picks one of three execution modes (workflow scripts, persistent named agents with SendMessage, or one-shot sub-agents) and one of six team patterns. A /harness:evolve skill feeds usage feedback back into the generated harness.
Who it is for: Claude Code users who want multi-agent setups generated for their project
Strengths
- Generates plain files in .claude/agents and .claude/skills that you can edit and commit
- Covers six team patterns, each mapped to one of three execution modes
- Includes /harness:evolve and a mechanical migration path from v1 artifacts
- No environment variables or experimental flags required
Weaknesses
- Works only with Claude Code; no other agent host is mentioned
- Quality claims come from an author-run A/B test on 15 tasks
- Interactive walkthrough is in Korean only
- No release published; version 2.1.0 appears only in a badge
What it needs
- no GPU
- Needs Claude Code
- Models: Claude (fable, opus, sonnet tiers)
Also in Skills, plugins and rules
See all 85| Rank | Project | Score |
|---|---|---|
| 1 | skillsAgent skills for planning, TDD, debugging and code review | 94 out of 100 |
| 2 | eccAgents, skills, hooks and rules bundle for Claude Code, Codex and other harnesses | 94 out of 100 |
| 3 | cavemanCuts token use in coding agents with terse replies and compressed inputs | 94 out of 100 |
| #4 | ponytailSkill prompt that makes coding agents write less code and add tests | 92 out of 100 |
| #5 | impeccableDesign skill with 24 commands and a detector for AI-generated frontends | 92 out of 100 |