Token导航 LogoToken导航TokenDH.com
研究检索只读github未标认证来源可访问许可证需确认审计通过

conductor-newtrack指挥新轨道

Agent Skill

conductor-newtrack 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

225

周安装

9

GitHub Stars

2

下载量

73
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:conductor-newtrack(指挥新轨道)
来源仓库:https://github.com/airclear/skills
仓库路径:skills/conductor-newtrack
安装命令:
npx skills add https://github.com/airclear/skills --skill conductor-newtrack
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/airclear/skills --skill conductor-newtrack

简介

用于启动新任务前的规划阶段,确保每个工作项都有明确需求和执行路径。

  • 适合遵循“先规划后编码”原则的项目管理模式。
  • 自动生成 spec.md 和 plan.md 文件,约束后续开发行为。
  • 首次使用前需完成 Conductor 项目初始化设置,加载项目上下文。
  • conductor-newtrack 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

Conductor New Track

Overview

This skill facilitates the "Plan before you build" philosophy. It ensures that every task starts with clear requirements (spec.md) and a structured, actionable execution path (plan.md) that respects the project's established workflow. The entire track creation process runs within Plan Mode, providing a controlled environment for file creation and modification.

When to Use

  • When a user says "Start a new feature," "Fix a bug," or "I have a new task."
  • Before any implementation code is written.

Workflow

  1. Setup Check: Verify Conductor is set up and load project context (Product Definition, Tech Stack, Workflow).
  2. Enter Plan Mode: Call enter_plan_mode tool before proceeding. Get track description if not provided via args.
  3. Interactive Spec: Ask targeted questions to define success. Batch up to 4 related questions in a single ask_user tool call. Questions are classified as Additive (brainstorming/scope, multiSelect: true) or Exclusive Choice (foundational decisions, multiSelect: false). Draft spec.md and get user approval via embedded content review.
  4. Plan Generation: Create a phased to-do list based on the spec and the project's workflow.md. Inject Phase Completion Tasks if defined in the workflow. Draft plan.md and get user approval via embedded content review.
  5. Artifact Scaffolding: Create the track directory and initialize index, metadata, spec, and plan files. Ensure unique short names by checking existing tracks.
  6. Exit Plan Mode: Call exit_plan_mode tool with the path to the track's index.md.
  7. Registry Registration: Append the new track to the project's tracks.md in the format - [] **Track: <Description>** and commit the changes.

Implementation Details

Refer to the following protocols for detailed procedural instructions:

Mandatory Constraints

  • Plan Mode Execution: The entire newTrack process MUST run within Plan Mode. Enter via enter_plan_mode at the start and exit via exit_plan_mode after artifacts are created. Only write_file, replace, and authorized run_shell_command are permitted. All file paths MUST use conductor/ relative paths. Redirection (> or >>) is strictly prohibited in run_shell_command during Plan Mode.
  • Question Classification: Every question MUST be classified as either Additive (multiSelect: true, for brainstorming and scope definition) or Exclusive Choice (multiSelect: false, for foundational singular decisions). Each question MUST include: header (max 16 chars), type, multiSelect (for choice type), options with label and description, and placeholder (for text type).
  • Batch Questioning: Streamline the process by batching up to 4 related questions in a single ask_user tool call.
  • Workflow Alignment: Plans MUST include TDD tasks (Write Tests -> Implement) and Phase Completion meta-tasks if required by the project workflow.
  • Unique IDs: Verify that the track's short name is unique before generating the full ID.
  • Git Hygiene: Every new track creation MUST conclude with a commit of the tracks.md file.

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

35.26%
按下载量换算26

Claude

31.37%
按下载量换算23

Cursor

16.85%
按下载量换算12

Gemini CLI

8.39%
按下载量换算6

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills