Backlog-md 是 AI Skill Hub 本期精选Agent工作流之一。已获得 5.9k 颗 GitHub Star,综合评分 8.2 分,整体质量较高。我们强烈推荐将其纳入你的 AI 工具库,帮助提升工作效率。
Backlog-md 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。
Backlog-md 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:npm 全局安装 npm install -g backlog.md # 方式二:npx 直接运行(无需安装) npx backlog.md --help # 方式三:项目依赖安装 npm install backlog.md # 方式四:从源码运行 git clone https://github.com/MrLesk/Backlog.md cd Backlog.md npm install npm start
# 命令行使用
backlog.md --help
# 基本用法
backlog.md [options] <input>
# Node.js 代码中使用
const backlog.md = require('backlog.md');
const result = await backlog.md.run(options);
console.log(result);
# backlog.md 配置说明 # 查看配置选项 backlog.md --config-example > config.yml # 常见配置项 # output_dir: ./output # log_level: info # workers: 4 # 环境变量(覆盖配置文件) export BACKLOG.MD_CONFIG="/path/to/config.yml"
Markdown‑native Task Manager & Kanban visualizer for any Git repository
<p align="center"> <code>npm i -g backlog.md</code> or <code>bun add -g backlog.md</code> or <code>brew install backlog-md</code> or <code>nix run github:MrLesk/Backlog.md</code> </p>

---
Backlog.md turns any folder with a Git repo into a self‑contained project board powered by plain Markdown files and a zero‑config CLI. Built for spec‑driven AI development — structure your tasks so AI agents deliver predictable results.
.md filebacklog board paints a live board in your shellbacklog browser launches a sleek web UI for visual task managementbacklog searchbacklog board export creates shareable markdown reports---
```bash
bun i -g backlog.md
<details> <summary><strong>Claude Code</strong></summary>
claude mcp add backlog --scope user -- backlog mcp start
</details>
<details> <summary><strong>Codex</strong></summary>
codex mcp add backlog -- backlog mcp start
</details>
<details> <summary><strong>Gemini CLI</strong></summary>
gemini mcp add backlog -s user backlog mcp start
</details>
<details> <summary><strong>Kiro</strong></summary>
kiro-cli mcp add --scope global --name backlog --command backlog --args mcp,start
</details>
Use the shared backlog server name everywhere. The server finds the active project from your client's MCP roots, and re-resolves when you switch workspace or worktree. Until it finds one, it serves backlog://init-required. A single user-scope server covers every repo.
{
"mcpServers": {
"backlog": {
"command": "backlog",
"args": ["mcp", "start"],
"env": {
"BACKLOG_CWD": "/absolute/path/to/your/project"
}
}
}
}
Set BACKLOG_CWD to pin the server to one project and stop workspace following. Use it to always target the same backlog, or when your client can't report MCP roots. If your IDE supports custom args but not env vars, you can also use ["mcp", "start", "--cwd", "/absolute/path/to/your/project"].
[!IMPORTANT] When adding the MCP server manually, add a short instruction to your CLAUDE.md/AGENTS.md files telling agents to readbacklog://workflow/overview. This step is not required when usingbacklog initas it adds these instructions automatically. For CLI-based setups, usebacklog instructions overviewto fetch the current workflow guidance.
Once connected, agents can read the Backlog.md workflow instructions via backlog://workflow/overview, with detailed guides at backlog://workflow/task-creation, backlog://workflow/task-execution, and backlog://workflow/task-finalization. Use /mcp command in your AI tool (Claude Code, Codex, Kiro) to verify if the connection is working.
---
Backlog.md merges the following layers (highest → lowest):
1. CLI flags 2. Project config file: - backlog.config.yml when present - otherwise backlog/config.yml or .backlog/config.yml 3. Built‑ins
Run backlog config with no arguments to launch the full interactive wizard. This is the same experience triggered from backlog init when you opt into advanced settings, and it walks through the complete configuration surface: - Cross-branch accuracy: checkActiveBranches, remoteOperations, and activeBranchDays. - Git workflow: autoCommit and bypassGitHooks. - ID formatting: enable or size zeroPaddedIds. - Editor integration: pick a defaultEditor with availability checks. - Definition of Done defaults: interactively add/remove/reorder/clear project-level definition_of_done checklist items. - Web UI defaults: choose defaultPort and whether autoOpenBrowser should run.
Skipping the wizard (answering "No" during init) applies the safe defaults that ship with Backlog.md: - checkActiveBranches=true, remoteOperations=true, activeBranchDays=30. - autoCommit=false, bypassGitHooks=false. - zeroPaddedIds disabled. - defaultEditor unset (falls back to your environment). - defaultPort=6420, autoOpenBrowser=true.
For filesystem-only projects, run backlog init --no-git. Backlog.md will not run git init, and the saved config forces checkActiveBranches=false, remoteOperations=false, and autoCommit=false so CLI, Web, and MCP local-file workflows do not depend on a Git repository.
Whenever you revisit backlog init or rerun backlog config, the wizard pre-populates prompts with your current values so you can adjust only what changed.
Launch a modern, responsive web interface for visual task management:
```bash
Full command reference — task management, search, board, docs, decisions, and more: CLI-INSTRUCTIONS.md
Quick examples: backlog, backlog instructions, backlog task create, backlog task list, backlog task edit, backlog milestone add, backlog milestone rename, backlog milestone remove, backlog search, backlog board, backlog browser.
Full help: backlog --help
---
CLI instructions are the default AI setup. MCP remains supported for AI coding assistants like Claude Code, Codex, Gemini CLI and Kiro when you explicitly prefer an MCP connector. You can run backlog init (even if you already initialized Backlog.md) and choose MCP integration, or follow the manual steps below.
高质量的开源AI工作流管理工具
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。
经综合评估,Backlog-md 在Agent工作流赛道中表现稳健,质量优秀。如果你已有明确的使用需求,可以直接上手体验;如果还在评估阶段,建议对比同类工具后再做决策。
| 原始名称 | Backlog-md |
| Topics | workflowagentagentic-aimanagementmarkdown |
| GitHub | https://github.com/MrLesk/Backlog.md |
| License | MIT |
| 语言 | TypeScript |
收录时间:2026-07-01 · 更新时间:2026-07-01 · 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