AI Skill Hub 推荐使用:AI 工作流工具 是一款优质的Agent工作流。AI 综合评分 7.5 分,在同类工具中表现稳健。如果你正在寻找可靠的Agent工作流解决方案,这是一个值得深入了解的选择。
AI 工作流工具 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
AI 工作流工具 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
# 方式一:pip 安装(推荐)
pip install harness-kit
# 方式二:虚拟环境安装(推荐生产环境)
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install harness-kit
# 方式三:从源码安装(获取最新功能)
git clone https://github.com/misty-step/harness-kit
cd harness-kit
pip install -e .
# 验证安装
python -c "import harness_kit; print('安装成功')"
# 命令行使用
harness-kit --help
# 基本用法
harness-kit input_file -o output_file
# Python 代码中调用
import harness_kit
# 示例
result = harness_kit.process("input")
print(result)
# harness-kit 配置文件示例(config.yml) app: name: "harness-kit" debug: false log_level: "INFO" # 运行时指定配置文件 harness-kit --config config.yml # 或通过环境变量配置 export HARNESS_KIT_API_KEY="your-key" export HARNESS_KIT_OUTPUT_DIR="./output"
29 catalog skills, 11 core agents, and harness infrastructure for AI-assisted software development. One repo. All harnesses (Claude Code, Codex, Pi, Antigravity).
Harness Kit is an operator-facing harness primitive library, not a buyer-facing governed workflow package or admin-control plane. Read docs/positioning.md before framing this repo for clients, departments, procurement, security reviewers, or executives.
```bash
| Skill | Purpose |
|---|---|
/deliver | Inner-loop composer: ticket → merge-ready (shape → implement → review+ci+refactor+qa) |
/flywheel | Outer-loop orchestrator: cycles of /deliver → /monitor → /reflect |
/code-review | Parallel multi-agent review, auto-fix loop |
/diagnose | Investigate, triage, fix |
/qa | Verify the changed surface and capture evidence |
/hardening | Property tests, mutation testing, CRAP/SCRAP, DRY, and acceptance mutation |
/demo | Show what changed with the right artifact for the change shape |
/design | Artifact-backed critique and polish for hierarchy, typography, layout, and taste |
/monitor | Watch post-change signals and escalate regressions |
/groom | Backlog management, brainstorming, rethink, scaffold |
/harness-engineering | Skill engineering, primitive management, context lifecycle |
/create-repo-skill | Generate repo-local QA, persona acceptance, and bespoke workflow skills |
/reflect | Session retrospective, harness postmortem, operator coaching |
/research | Multi-source web research, delegation, think tank |
/shape | Spec/design → context packet output |
backlog.d/ → /groom → /shape → /deliver → ship
└─ /flywheel (outer loop cycles:
/deliver → /monitor → /reflect → next)
For a deeper map of the repository architecture, encoded assumptions, operating loop, and active backlog, read CODEBASE.md.
高质量的AI工作流工具,功能丰富
AI Skill Hub 为第三方内容聚合平台,本页面信息基于公开数据整理,不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后,再部署至生产环境,并做好必要的安全评估。
✅ MIT 协议 — 最宽松的开源协议之一,可自由商用、修改、分发,仅需保留版权声明。
总体来看,AI 工作流工具 是一款质量良好的Agent工作流,在同类工具中具备一定竞争力。AI Skill Hub 将持续追踪其更新动态,建议收藏备用,结合自身场景选择合适时机引入使用。
| 原始名称 | harness-kit |
| 原始描述 | 开源AI工作流:Portable skill library for AI coding agents: debugging, PR workflows, design sys。⭐13 · Python |
| Topics | AI工作流Python |
| GitHub | https://github.com/misty-step/harness-kit |
| License | MIT |
| 语言 | Python |
收录时间:2026-06-04 · 更新时间:2026-06-04 · License:MIT · AI Skill Hub 不对第三方内容的准确性作法律背书。
选择 Agent 类型,复制安装指令后粘贴到对应客户端