Agent skill

Subagent Coordinator

by flyxl in flyxl/datazen

Orchestrate multi-track parallel feature development with subagents and git worktrees.

GPL-3.0Auto-check passedAgent Workflows

Install Subagent Coordinator

skills CLI
$ npx skills add flyxl/datazen --skill subagent-coordinator -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install flyxl/datazen subagent-coordinator --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/flyxl/datazen.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/subagent-coordinator .claude/skills/subagent-coordinator && rm -rf skills-src

Use ~/.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/

Facts

Skill name
subagent-coordinator
GitHub stars
114
Token cost
~908 tokens
SKILL.md length
253 words
Files
1
Skills in repo
4
Repo updated
First seen
Licence
GPL-3.0

At a glance

Orchestrate multi-track parallel feature development with subagents and git worktrees.

  • Works in 7 steps: 波次与冲突面拆解 → 准备并行轨道 (Bootstrap) → 并行派发编码子代理 (Coder) → …
  • Executing complex multi-feature tasks
  • SKILL.md covers 核心硬性纪律, 标准协调工作流 and 简报七件(质量的最大变量)
  • Calls git, pnpm and npx

What it does

Subagent Coordinator is an agent skill from flyxl/datazen. Orchestrate multi-track parallel feature development with subagents and git worktrees. Use when executing complex multi-feature tasks, large refactorings, or other work that benefits from parallel coding/testing agents in DataZen.

Its SKILL.md is about 910 tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Agent Workflows, covering Subagents, Git worktrees and Refactoring. The repository describes itself as: Lightweight GPL-3.0 desktop database client (PostgreSQL / MySQL / SQLite / Redis). Built with Tauri + Rust, under 15MB. The licence is GPL-3.0.

When your agent uses it

  • Executing complex multi-feature tasks
  • Large refactorings
  • Other work that benefits from parallel coding/testing agents in DataZen

Example prompts

  • “/subagent-coordinator”

Requirements

  • Node.js

Workflow steps

7 steps, taken from the step headings in SKILL.md.

  1. 波次与冲突面拆解
  2. 准备并行轨道 (Bootstrap)
  3. 并行派发编码子代理 (Coder)
  4. 即时测试派发 (Tester)
  5. 活性监控与断点恢复
  6. 合流与 Worktree 清理
  7. 全量回归 (R 阶段)

What it can do on your machine

Read from SKILL.md and the folder at commit 0ed2e50. 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
    • pnpm
    • npx
    • cargo

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use git, pnpm and npx, 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

Subagent Coordinator loads about 908 tokens when it runs. Until then it costs about 63 tokens; SKILL.md has 253 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~63
When it runs · the whole SKILL.md, loaded when a task matches
~908

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 flyxl/datazen at commit 0ed2e50, republished under its GPL-3.0 licence (© flyxl). 253 words, ~908 tokens.

Download SKILL.mdSave it as .claude/skills/subagent-coordinator/SKILL.md (or your agent's skills folder).
name
subagent-coordinator
description
Orchestrate multi-track parallel feature development with subagents and git worktrees. Use when executing complex multi-feature tasks, large refactorings, or other work that benefits from parallel coding/testing agents in DataZen.

Subagent Parallel Development Coordinator

本 Skill 指导主会话 Agent 作为协调者 (Coordinator),负责统筹、拆轨、派发并合流多子代理并行开发与测试任务。

不写台账。 进度、缺陷清单、设计方案一律不落文件。结论必须进代码、测试与 docs/ 的正式文档(docs/features/、docs/architecture/、docs/development/)。过程产物(如需临时记录)只存在于会话上下文,随会话消亡。 完整项目约定见 AGENTS.md。

核心硬性纪律

  1. 协调者不亲自写大段业务代码:专注架构分解、简报编写、状态推进与合流裁决。
  2. 严防主线污染:所有开发均在 .worktrees/datazen-<track-id> 中进行,主检出仅用于合流与清理。
  3. 不派发不存在的简报模板:简报由协调者现场编写(见 §简报七件),仓库内没有现成模板文件。

标准协调工作流

1. 波次与冲突面拆解
  • 依据文件冲突面划分子任务轨道,而非功能相邻度。
  • 触碰互斥文件或同一注册块不同行可并行(同一 Wave);存在核心依赖或模式复用软依赖必须串行(分 Wave)。
  • 确定基准集成分支(如 feat/<initiative-name>)。
2. 准备并行轨道 (Bootstrap)

对于当前波次的每个 <track-id>,在主检出执行:

bash
scripts/new-feature-worktree.sh <track-id> <base-branch>

脚本会自动在 .worktrees/datazen-<track-id> 初始化分支、软链依赖、代码生成,并拷贝本地未跟踪文档(含 gitignored 的 posts/)。

3. 并行派发编码子代理 (Coder)
  • 硬性要求:同一 Wave 的全部 Coder 必须在同一条消息中通过多个 subagent 工具调用并行派发,严禁串行逐个启动。
  • 每个 Coder 启动 全新实例。
  • 必读清单必须包含:AGENTS.md + 该轨相关的正式文档 + 已侦察落点。
  • 明确工作目录 pwd 为 .worktrees/datazen-<track-id>。
  • 返回格式约定:commit hash / 改动清单 / 各套件实测数字 / 遗留项,状态词用 READY_FOR_TEST。
4. 即时测试派发 (Tester)
  • 硬性要求:某个 Coder 返回 READY_FOR_TEST 后,协调者必须立即为该轨启动 Tester,无需等待同 Wave 其他 Coder 完成。
  • 若多个 Coder 同时返回,Tester 也必须在同一条消息中并行派发。
  • 启动 全新测试代理(禁止复用编码代理——它有立场)。
  • Tester 职责不仅是重跑编码代理的自测,还必须:
    1. Review 实现逻辑——审查代码变更是否正确、是否有边界遗漏。
    2. 覆盖率驱动的测试补齐——确保改动代码覆盖率达标,精准补齐未覆盖的分支和路径,不追求数量。
    3. 设计 E2E 测试用例——直接编写可执行的 E2E 测试,不登记待办。
  • 测试代理只测不修。缺陷逐条写入最终报告(ID + 描述 + 复现步骤 + 日志),不落盘单文件。
4.1 Bug 修复循环

Tester 完成一轮完整测试后统一上报缺陷清单 + TEST_FAILED,协调者启动修复循环:

text
Tester 完成完整测试 → 上报缺陷清单 + TEST_FAILED
→ 协调者指派原 Coder agent(resume)修复
→ Coder 修复并提交 → 协调者派发全新 Tester 完整复测
→ 通过 → 闭环 / 不通过 → 回到循环起点

硬性规则:

  1. 完整上报:Tester 在完成全部测试阶段后统一上报,不逐个打断。
  2. 复用原 Coder:修复阶段优先 resume 原编码 agent,利用其已有上下文。仅当原 Coder 不可恢复时才启动全新 Rescuer。
  3. 修复后必须复测:Coder 的自验不能替代 Tester 复测,禁止跳过复测直接合入。
  4. 全新 Tester:每轮复测必须使用全新 Tester 实例。
  5. 最大循环次数:同一轨道最多 5 轮完整修复-复测循环。超过仍有未修复缺陷,向用户汇报并由用户决定下一步。
  6. 修复 Coder 简报:必须包含完整缺陷清单,以及「仅修复这些,不做范围外改动」的纪律约束。
  7. 范围归属裁决:既有缺陷是否并入本功能范围由协调者裁决——默认不并入已关闭功能,另立循环。
5. 活性监控与断点恢复
  • 编码代理给予 20 分钟纯探索宽限期。
  • 依据 git -C <worktree> status --short、系统进程和构建产物 mtime 判定活性。
  • 死亡恢复:
    • 死亡 ≤3 次:发消息 "继续" 原实例续跑(保留完整上下文与已有编辑,远优于新实例)。
    • 死亡 >3 次:启动全新接管代理 (Rescuer):先用 git status 盘点现场未提交改动 → 审计已有 diff → 补缺口 → 验证 → 提交。
6. 合流与 Worktree 清理

当某轨测试通过闭环后,协调者在集成分支合流:

bash
# 1. 合并功能分支
git merge --no-ff feature/<track-id> -m "feat(<track-id>): merge"

# 2. 健全性校验
pnpm typecheck
npx vitest run

# 3. 释放 worktree 及分支
git worktree remove .worktrees/datazen-<track-id>
git branch -d feature/<track-id>
7. 全量回归 (R 阶段)

所有 Wave 的全部轨道合并完毕后,统一执行一次 R 阶段全量回归(中间各 Wave 合入时仅做合并健全性校验):

  1. 跑通全套构建与测试:pnpm typecheck、npx vitest run、cargo test -p datazen --lib、必要的 E2E。
  2. 逐项回归各轨在简报中标记为【留待 R 回归】的 E2E 用例。
  3. 确认全部缺陷已闭环。

简报七件(质量的最大变量)

  1. 工作目录与禁区:worktree 绝对路径;明确禁止触碰主检出与其他轨道。
  2. 必读清单:项目约定(AGENTS.md)+ 相关正式文档,按序。
  3. 任务与验收标准:可逐条判定的完成标准,不接受模糊表述。
  4. 已侦察落点:协调者预先 grep 好的 文件:行——这是代理效率差异的最大变量;并声明「落点需自行核实」防盲从。
  5. 执行纪律:一律用 Grep 工具搜索(禁 bash 全仓 grep);编辑前先 read(edit 工具强制要求);文件被 git 操作或他人改动后必须重读。
  6. 环境注意:node_modules 是软链,禁 pnpm install,二进制用 npx;并行 cargo 轨各自设 CARGO_TARGET_DIR。
  7. 返回格式:commit hash / 改动清单 / 各套件数字 / 遗留。

© flyxl, GPL-3.0. 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 .agents/skills/subagent-coordinator of flyxl/datazen.

Open the folder on GitHubat commit 0ed2e50

Compare with similar skills

Subagent Coordinator 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.

Subagent Coordinator compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Subagent Coordinator this skillflyxl/datazen114—~908Automated safety check: PassGPL-3.0
Batch Orchestrationrohitg00/pro-workflow2.9k—~1.2kAutomated safety check: PassNone
Codex CLIkortix-ai/suna20k—~1.6kAutomated safety check: PassCustom licence
Antigravity Agentsmarkfulton/claude-antigravity-agents130—~2.1kAutomated safety check: PassMIT
Vspawnvlinx-io/VelaTerm259—~2.5kAutomated safety check: PassMIT
Spec-Driven Development v2LichAmnesia/lich-skills234—~3.1kAutomated safety check: PassMIT

Similar skills

  • Batch Orchestration

    rohitg00/pro-workflow

    Decompose large-scale changes into independent units and spawn parallel agents in isolated worktrees.

    2.9k GitHub stars~1.2k tokensUpdated 8 days ago
    Agent WorkflowsAuto-check passed
  • Codex CLI

    kortix-ai/suna

    Drive OpenAI's Codex CLI (codex exec) as a non-interactive coding sub-agent from inside Claude Code.

    20k GitHub stars~1.6k tokensUpdated today
    Agent WorkflowsAuto-check passed
  • Antigravity Agents

    markfulton/claude-antigravity-agents

    Delegate coding, code review, analysis, and research jobs to Google Antigravity CLI (agy) sub-agents that run alongside your own work.

    130 GitHub stars~2.1k tokensUpdated 2 mo ago
    Agent WorkflowsAuto-check passed
  • Vspawn

    vlinx-io/VelaTerm

    Explicitly spawn a standalone child session under the current vlx-term session, passing the task in as its first message (mirrors spawntask).

    259 GitHub stars~2.5k tokensUpdated today
    Agent WorkflowsAuto-check passed
  • Spec-Driven Development v2

    LichAmnesia/lich-skills

    Organizes long-running agent work into a Project, Sprint and Task hierarchy with per-task state files, isolated worktrees, review loops and script-checked rules.

    234 GitHub stars~3.1k tokensUpdated 3 mo ago
    Agent WorkflowsAuto-check passed
  • Plan and execute large refactor or rewrite efforts efficiently with parallel multi-agent analysis and implementation.

    4k GitHub stars~889 tokensUpdated 6 mo ago
    Agent WorkflowsAuto-check passed

More from flyxl/datazen

  • Bump Version

    flyxl/datazen

    Update the DataZen application version number across all files and trigger a GitHub release rebuild.

    114 GitHub stars~392 tokensUpdated 7 days ago
    Auto-check passed
  • Gen Ctx YAML

    flyxl/datazen

    Generate .ctx.yaml table group context files for NL2SQL optimization.

    114 GitHub stars~661 tokensUpdated 7 days ago
    Auto-check passed
  • Wapp

    flyxl/datazen

    Create, scaffold, modify, debug, package and test a DataZen Workspace App (Wapp / UI Extension), a sandboxed runtime app that contributes a workspace page and/or themes.

    114 GitHub stars~1.9k tokensUpdated 7 days ago
    Auto-check passed

Questions about Subagent Coordinator

What does Subagent Coordinator do?

Orchestrate multi-track parallel feature development with subagents and git worktrees. Subagent Coordinator is an agent skill from flyxl/datazen. Orchestrate multi-track parallel feature development with subagents and git worktrees.

When should I use Subagent Coordinator?

Subagent Coordinator fits situations like: executing complex multi-feature tasks; large refactorings; other work that benefits from parallel coding/testing agents in DataZen.

How do I install Subagent Coordinator in Claude Code?

Run `npx skills add flyxl/datazen --skill subagent-coordinator -a claude-code`. Or copy the skill folder (.agents/skills/subagent-coordinator in flyxl/datazen) into .claude/skills/subagent-coordinator in your project. Claude Code loads it when a task matches its description.

How do I install Subagent Coordinator in Codex?

Run `npx skills add flyxl/datazen --skill subagent-coordinator -a codex`. Or copy the skill folder (.agents/skills/subagent-coordinator in flyxl/datazen) into .agents/skills/subagent-coordinator in your project. Codex loads it when a task matches its description.

Can I use Subagent Coordinator 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 flyxl/datazen --skill subagent-coordinator -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/subagent-coordinator, .gemini/skills/subagent-coordinator, .github/skills/subagent-coordinator and .opencode/skills/subagent-coordinator in your project.

What does Subagent Coordinator need to run?

Going by SKILL.md and its folder, Subagent Coordinator needs the command-line tools its instructions call (git, pnpm, npx and cargo). Our summary lists: Node.js.

Does Subagent Coordinator access the network?

SKILL.md contains no URLs. Its commands use git and npx, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Subagent Coordinator 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 Subagent Coordinator use?

Subagent Coordinator is published under the GPL-3.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Subagent Coordinator use?

About 908 tokens (SKILL.md is roughly 3.6k 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 Subagent Coordinator?

Skills that share tags, products or a category with Subagent Coordinator: Batch Orchestration (rohitg00/pro-workflow, 2.9k stars), Codex CLI (kortix-ai/suna, 20k stars), Antigravity Agents (markfulton/claude-antigravity-agents, 130 stars) and Vspawn (vlinx-io/VelaTerm, 259 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Subagent Coordinator?

flyxl (a GitHub user) maintains it in flyxl/datazen, which has 114 GitHub stars. The repository holds 4 skills in this directory. The repository was last updated on September 30, 2026.

Source: flyxl/datazen on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.