Squad Agent Collaboration Patterns
microsoft/waza
Shared collaboration rules for a team of squad agents covering worktree awareness, writing decisions to an inbox, cross-agent requests and reviewer lockout.
Collects feedback on how an agent harness performed, generalizes it, and updates the harness agents, skills and orchestrator along with a change-history table.
SKILL.md written in Korean; this summary is our English description.
Install Harness Evolution Feedback Loop
$ npx skills add revfactory/harness --skill evolve -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install revfactory/harness evolve --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/revfactory/harness.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/evolve .claude/skills/evolve && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "evolve" agent skill from https://github.com/revfactory/harness/tree/main/skills/evolve into .claude/skills/evolve/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "evolve", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/revfactory/harness/tree/main/skills/evolveType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add revfactory/harness --skill evolve -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install revfactory/harness evolve --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/revfactory/harness.git skills-src && mkdir -p .agents/skills && cp -r skills-src/skills/evolve .agents/skills/evolve && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "evolve" agent skill from https://github.com/revfactory/harness/tree/main/skills/evolve into .agents/skills/evolve/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "evolve", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add revfactory/harness --skill evolve -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install revfactory/harness evolve --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/revfactory/harness.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/skills/evolve .cursor/skills/evolve && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "evolve" agent skill from https://github.com/revfactory/harness/tree/main/skills/evolve into .cursor/skills/evolve/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "evolve", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/revfactory/harness.git --path skills/evolve--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add revfactory/harness --skill evolve -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install revfactory/harness evolve --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/revfactory/harness.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/skills/evolve .gemini/skills/evolve && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "evolve" agent skill from https://github.com/revfactory/harness/tree/main/skills/evolve into .gemini/skills/evolve/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "evolve", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install revfactory/harness evolveInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add revfactory/harness --skill evolve -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/revfactory/harness.git skills-src && mkdir -p .github/skills && cp -r skills-src/skills/evolve .github/skills/evolve && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "evolve" agent skill from https://github.com/revfactory/harness/tree/main/skills/evolve into .github/skills/evolve/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "evolve", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add revfactory/harness --skill evolve -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install revfactory/harness evolve --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/revfactory/harness.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/skills/evolve .opencode/skills/evolve && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "evolve" agent skill from https://github.com/revfactory/harness/tree/main/skills/evolve into .opencode/skills/evolve/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "evolve", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
evolve- GitHub stars
- 9.1k
- Token cost
- ~855 tokens
- SKILL.md length
- 517 words
- Files
- 1
- Skills in repo
- 2
- Repo updated
- First seen
- Licence
- Apache-2.0
At a glance
Collects feedback on how an agent harness performed, generalizes it, and updates the harness agents, skills and orchestrator along with a change-history table.
- Works in 5 steps: 델타 수집 → 피드백 유형 분류 및 수정 대상 매핑 → 일반화 및 반영 → …
- Folding feedback on a disappointing harness result back into its skills
- SKILL.md covers 워크플로우 and 원칙
- Calls git
What it does
The skill is written in Korean for an existing agent harness, a setup of agents, skills and an orchestrator, and treats that harness as something that evolves. It gathers the difference between the initial configuration and current use by reading `.claude/agents/`, `.claude/skills/` and the change-history table in CLAUDE.md, checking git history of those files and scanning the `_workspace/` folder for traces of recent runs. It asks you for feedback but proposes improvements on its own when it sees repeated fix requests, repeated agent failures or manual workarounds that bypass the orchestrator.
Feedback is classified and mapped to a target: output quality goes to a skill, a role problem to an agent definition, ordering to the orchestrator, missed triggers to a skill description, and cost or scale problems to the orchestrator. Changes are generalized to principles rather than patched for one case, written with their reasons, applied one at a time, and checked against earlier history so a fix does not undo a previous one. Each change is logged in CLAUDE.md and the edited files are validated, with trigger tests when a description changes. Adding agents or redesigning the structure is left to the separate harness skill.
When your agent uses it
- Folding feedback on a disappointing harness result back into its skills
- Reviewing how a harness has changed since it was first set up
- Fixing a skill description that fails to trigger on a phrase
- Recording a harness change in the CLAUDE.md change history
Example prompts
- “The report from the last harness run was too shallow. Update the harness so the next run goes deeper.”
- “Review what changed in my .claude agents and skills since the first setup and suggest improvements.”
- “This phrase did not trigger my orchestrator skill. Extend its description and log the change.”
Requirements
- An existing harness with `.claude/agents/`, `.claude/skills/` and a CLAUDE.md change-history table
Workflow steps
5 steps, taken from the step headings in SKILL.md.
What it can do on your machine
Read from SKILL.md and the folder at commit 92d9f1b. It shows what the files ask for, not the result of running them.
Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
Shell commands in SKILL.md call:
git
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md. Its commands use git, which can reach the network depending on how they are called.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Context cost
Harness Evolution Feedback Loop loads about 855 tokens when it runs. Until then it costs about 72 tokens; SKILL.md has 517 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
Safety
Auto-check passed
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
SKILL.md
The full file from revfactory/harness at commit 92d9f1b, republished under its Apache-2.0 licence (© revfactory). 517 words, ~855 tokens.
Download SKILL.mdSave it as
.claude/skills/evolve/SKILL.md (or your agent's skills folder).- name
- evolve
- description
- 하네스 진화 스킬. 사용 중인 하네스의 실행 결과에 대한 피드백을 수집·일반화하여 에이전트/스킬/오케스트레이터에 반영하고, 초기 구성 대비 델타를 포착해 변경 이력을 갱신한다. '하네스 회고', '하네스 진화', '하네스 피드백 반영', '하네스 개선', '결과가 아쉬웠어 하네스 고쳐줘', '이 피드백 하네스에 반영해줘', '하네스 레슨 정리' 등 기존 하네스의 실행 경험을 바탕으로 한 개선 요청 시 반드시 이 스킬을 사용. 하네스 신규 구축·구조 재설계·에이전트 추가는 harness 스킬이 담당.
Harness Evolve — 하네스 진화 메커니즘
하네스는 고정물이 아니라 진화하는 시스템이다. 이 스킬은 "무엇이 먹혔고 무엇이 안 먹혔는가"의 델타를 포착해 하네스에 되먹여, 다음 실행이 측정 가능하게 더 나아지도록 한다.
초기 하네스 ──▶ 실 프로젝트 사용 ──▶ 현재 하네스
│
▼ (evolve로 델타 포착)
피드백 일반화 → 에이전트·스킬·오케스트레이터 반영
│
▼
변경 이력 갱신 → 다음 실행은 더 나은 초안에서 시작워크플로우
Phase 1: 델타 수집
.claude/agents/,.claude/skills/,CLAUDE.md(변경 이력 테이블)를 읽는다- git 저장소라면 하네스 파일들의 변경 이력을 조회한다 (
git log --oneline -- .claude/ CLAUDE.md) — 초기 구성 대비 무엇이 언제 왜 바뀌었는지 파악 _workspace/가 있으면 최근 실행의 중간 산출물을 훑어 실제 실행 흔적을 확인한다:- 오케스트레이터가 정의한 경로에 산출물이 실제로 있는가 (없으면 워크플로우가 우회되었거나 죽은 코드)
- 산출물 품질이 스킬이 명시한 형식/기준을 따르는가
- 사용자에게 피드백을 요청한다 (이미 피드백을 제공했다면 생략):
- "결과에서 개선할 부분이 있나요?"
- "에이전트 구성이나 워크플로우에 바꾸고 싶은 점이 있나요?"
- 피드백이 없으면 강요하지 않는다. 단, 아래 관찰 신호가 있으면 선제적으로 개선을 제안한다
관찰 기반 진화 신호 (피드백이 없어도 제안):
- 같은 유형의 수정 요청이 2회 이상 반복된 흔적
- 에이전트가 반복적으로 실패/재시도한 패턴
- 사용자가 오케스트레이터를 우회해 수동으로 작업한 흔적 (오케스트레이터 트리거 실패 의심 → description 확장 후보)
- 오케스트레이터에 v1 유물(TeamCreate/TeamDelete/실험 플래그)이 남아 있음 → harness 스킬의 마이그레이션 절차 안내
Phase 2: 피드백 유형 분류 및 수정 대상 매핑
| 피드백 유형 | 수정 대상 | 예시 |
|---|---|---|
| 결과물 품질 | 해당 에이전트의 스킬 | "분석이 너무 피상적" → 스킬에 깊이 기준 추가 |
| 에이전트 역할 | 에이전트 정의 .md | "보안 검토도 필요" → harness 스킬로 에이전트 추가 안내 |
| 워크플로우 순서 | 오케스트레이터 스킬 | "검증을 먼저 해야" → Phase 순서 변경 |
| 팀 구성 | 오케스트레이터 + 에이전트 | "이 둘은 합쳐도 될 듯" → 에이전트 병합 |
| 트리거 누락 | 스킬 description | "이 표현으로 하면 작동 안 함" → description 확장 |
| 실행 모드 부적합 | 오케스트레이터 | "매번 같은 팬아웃인데 느려" → 워크플로우 모드 전환 |
| 규모/비용 | 오케스트레이터 | "토큰을 너무 써" → 기본 규모 축소, 버짓 연동 추가 |
범위 판단: 에이전트 신규 추가/삭제나 아키텍처 재설계가 필요하면 이 스킬에서 직접 하지 않고 harness 스킬(0단계의 기존 구성 확장 절차)로 안내한다. evolve는 기존 구성의 조정에 집중한다.
Show full SKILL.md (234 more words)Show less
Phase 3: 일반화 및 반영
- 피드백을 일반화한다 — 특정 사례에만 맞는 좁은 수정은 오버피팅이다. "이번 보고서에 서론이 길었다" → "서론은 전체의 10% 이내로"가 아니라, 왜 길어졌는지(스킬에 분량 배분 기준 부재)를 찾아 원리 수준으로 수정한다
- Why를 함께 기록한다 — 수정된 지시에는 이유를 병기한다. 이유를 알면 에이전트가 엣지 케이스에서도 올바르게 판단한다
- 변경은 한 번에 하나씩 적용하고, 각 변경 직후 Phase 4를 실행한다
- 퇴행 방지: 수정이 기존 변경 이력의 이전 수정을 되돌리는 방향이면, 사용자에게 상충을 알리고 확인받는다 (과거에 "너무 길다"로 줄였는데 이번에 "너무 짧다"면 — 둘 다 만족하는 기준을 찾는 것이 정답이다)
- 기존 파일의 언어를 유지한다 — 에이전트·스킬·오케스트레이터·
CLAUDE.md에 반영하는 문장은 수정 대상 파일에 이미 쓰인 언어로 쓴다. 이 스킬 문서가 한국어라는 이유로 다른 언어로 된 하네스에 한국어 문장을 섞지 않는다
Phase 4: 변경 이력 갱신 및 검증
- CLAUDE.md의 변경 이력 테이블에 기록한다:
**변경 이력:**
| 날짜 | 변경 내용 | 대상 | 사유 |
|------|----------|------|------|
| 2026-07-19 | 톤 가이드 추가 | skills/content-creator | "너무 딱딱하다" 피드백 |- 수정된 파일의 구조를 검증한다 (frontmatter, 참조 일관성)
- description을 수정했다면 트리거 검증 (should-trigger + near-miss 각 3개 이상)
- CLAUDE.md와 실제 파일의 일치 여부 최종 확인
Phase 5: 진화 보고
사용자에게 보고한다:
- 포착된 델타 요약 (초기 구성 → 현재)
- 이번에 반영한 변경과 그 일반화 근거
- 반영하지 않기로 한 피드백과 이유 (있다면)
- 다음 실행에서 기대되는 개선점
원칙
- 델타는 자산이다 — 변경 이력이 쌓일수록 같은 도메인의 다음 하네스 구축이 "출시 상태에 더 가까운 초안"에서 시작된다. 이력을 지우지 않는다
- 일반화 없는 반영 금지 — 사례 하나에 규칙 하나를 1:1로 추가하면 스킬이 규칙 더미가 된다. 원리로 압축한다
- 한 번에 하나씩 — 여러 변경을 몰아서 적용하면 어떤 변경이 효과였는지 알 수 없게 된다
© revfactory, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Files
Just SKILL.md in skills/evolve of revfactory/harness.
Open the folder on GitHubat commit 92d9f1b
Compare with similar skills
Harness Evolution Feedback Loop next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Harness Evolution Feedback Loop this skillrevfactory/harness | 9.1k | — | ~855 | Automated safety check: Pass | Apache-2.0 | |
| Squad Agent Collaboration Patternsmicrosoft/waza | 1.4k | 4 repos | ~500 | Automated safety check: Pass | MIT | |
| Project KickoffStanshy/AgentHub | 201 | — | ~528 | Automated safety check: Notes | MIT | |
| Prompt Security Hardeninged3dai/ed3d-plugins | 250 | — | ~2.5k | Automated safety check: Warn | None | |
| Darwin Skill Optimizeralchaincyf/darwin-skill | 6.2k | 1 repos | ~4.7k | Automated safety check: Pass | MIT | |
| Neat-Freak Knowledge CloseoutKKKKhazix/khazix-skills | 21k | — | ~1.9k | Automated safety check: Pass | MIT |
Similar skills
- Official
Project Kickoff
Stanshy/AgentHub
Initialize new project with CLAUDE.md, .knowledge/ structure, company rules, and git
Prompt Security Hardening
ed3dai/ed3d-plugins
A skill your agent uses when writing skills, CLAUDE.md files, agent prompts, or any directives that involve shell commands, environment variables, API credentials, file creation, or git operations -…
Darwin Skill Optimizer
alchaincyf/darwin-skill
Scores SKILL.md files on a nine-dimension rubric, then improves them in a keep-or-revert loop with independent judge agents, test prompts, git history and human checkpoints.
Neat-Freak Knowledge Closeout
KKKKhazix/khazix-skills
Brings project docs, agent rule files, authorized memory and leftover workspace files back in line with what the code and runtime actually do at the end of a work session.
O2 Review Loop
openobserve/openobserve
Splits a change into planner, coder and independent reviewer roles: you confirm a spec, a subagent implements it, and a separate reviewer checks each round's local WIP commit.
More from revfactory/harness
Harness Agent Team Designer
revfactory/harness
Designs a project-specific agent harness: defines specialist agents, writes the skills they follow, picks an execution mode and model for each, and keeps the setup maintained.
Related
Works with
Categories
Questions about Harness Evolution Feedback Loop
What does Harness Evolution Feedback Loop do?
Collects feedback on how an agent harness performed, generalizes it, and updates the harness agents, skills and orchestrator along with a change-history table. The skill is written in Korean for an existing agent harness, a setup of agents, skills and an orchestrator, and treats that harness as something that evolves.md, checking git history of those files and scanning the `_workspace/` folder for traces of recent runs.
When should I use Harness Evolution Feedback Loop?
Harness Evolution Feedback Loop fits situations like: folding feedback on a disappointing harness result back into its skills; reviewing how a harness has changed since it was first set up; fixing a skill description that fails to trigger on a phrase; recording a harness change in the CLAUDE.md change history.
How do I install Harness Evolution Feedback Loop in Claude Code?
Run `npx skills add revfactory/harness --skill evolve -a claude-code`. Or copy the skill folder (skills/evolve in revfactory/harness) into .claude/skills/evolve in your project. Claude Code loads it when a task matches its description.
How do I install Harness Evolution Feedback Loop in Codex?
Run `npx skills add revfactory/harness --skill evolve -a codex`. Or copy the skill folder (skills/evolve in revfactory/harness) into .agents/skills/evolve in your project. Codex loads it when a task matches its description.
Can I use Harness Evolution Feedback Loop in Cursor, Gemini CLI or GitHub Copilot?
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add revfactory/harness --skill evolve -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/evolve, .gemini/skills/evolve, .github/skills/evolve and .opencode/skills/evolve in your project.
What does Harness Evolution Feedback Loop need to run?
Going by SKILL.md and its folder, Harness Evolution Feedback Loop needs the command-line tools its instructions call (git). Our summary lists: An existing harness with `.claude/agents/`, `.claude/skills/` and a CLAUDE.md change-history table.
Does Harness Evolution Feedback Loop access the network?
SKILL.md contains no URLs. Its commands use git, which can reach the network depending on how they are called. This is read from the text; nothing was executed.
Is Harness Evolution Feedback Loop safe to install?
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
What licence does Harness Evolution Feedback Loop use?
Harness Evolution Feedback Loop is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Harness Evolution Feedback Loop use?
About 855 tokens (SKILL.md is roughly 3.4k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
What are the alternatives to Harness Evolution Feedback Loop?
Skills that share tags, products or a category with Harness Evolution Feedback Loop: Squad Agent Collaboration Patterns (microsoft/waza, 1.4k stars), Project Kickoff (Stanshy/AgentHub, 201 stars), Prompt Security Hardening (ed3dai/ed3d-plugins, 250 stars) and Darwin Skill Optimizer (alchaincyf/darwin-skill, 6.2k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Harness Evolution Feedback Loop?
revfactory (a GitHub user) maintains it in revfactory/harness, which has 9,127 GitHub stars. The repository holds 2 skills in this directory. The repository was last updated on September 28, 2026.
Source: revfactory/harness on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.