Token导航 LogoToken导航

Agent Skill · 研究检索

todo-task-planning(待办事项计划)

todo-task-planning 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。

todo-task-planning 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为只读;许可证信息还不明确,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 55,255。

总安装 489周安装 21GitHub Stars 公开资料未说明下载 171核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

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:todo-task-planning(待办事项计划)
来源仓库:https://github.com/gendosu/agent-skills
仓库路径:skills/todo-task-planning
安装命令:
npx skills add https://github.com/gendosu/agent-skills --skill todo-task-planning
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/gendosu/agent-skills --skill todo-task-planning

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

todo-task-planning 用于查找、检索和筛选相关信息。适合在 Codex、Claude、Cursor、Gemini CLI 中根据关键词、任务场景或来源线索快速定位候选结果。通过 npx skills add 命令从指定 GitHub 仓库安装使用。

安装前需确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。建议结合原始 README 核验具体用法和功能边界。

来源仓库内路径:skills/todo-task-planning

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/gendosu/agent-skills --skill todo-task-plannin…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它标注为只读,先用只读任务试一次,确认行为符合预期再扩大使用范围。

常见问题

6 条

todo-task-planning 怎么安装?

在终端执行 npx skills add https://github.com/gendosu/agent-skills --skill todo-task-planning。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

todo-task-planning 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

todo-task-planning 需要什么权限?

权限级别是只读。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。

todo-task-planning 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

todo-task-planning 用的人多吗?

公开数据是累计安装 489 次,近一周安装 21 次,下载 171 次。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

todo-task-planning 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。
  • 标注为只读权限,试用阶段不容易误改本地文件。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。
  • 只想先试试效果、不希望它改动本地文件的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
todo-task-planning

待办事项计划 · 本页

489—只读未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Command arguments: $ARGUMENTS

Usage

/todo-task-planning <file_path> [--pr] [--branch <name>]

Arguments

  • file_path (required): Path to the file for task planning execution
  • --pr (optional): Create a pull request after task completion. When specified, tasks will include branch creation (auto-generated if --branch not specified), commits, and PR creation
  • --branch <name> (optional): Branch name to create and use for task execution. Creates the specified branch and commits all changes to it. Can be used independently or with --pr option

Command Overview

This command reads the specified file ($ARGUMENTS) and performs comprehensive task planning. It can be executed repeatedly on the same file and also manages, confirms, and updates questions.

Important: This command is designed to be executed repeatedly. Do not question the fact that the same command will be executed multiple times; analyze and plan with a fresh perspective each time.

Important: The file specified in $ARGUMENTS will be the TODO management file. Avoid including specifications and research results in this file as much as possible; use docs/memory to store research results. Also, do not excessively abbreviate research results; retain them. Check each time whether you have researched something in the past. Do not neglect checking to avoid duplicating research results and tasks.

Branch and PR Options Usage

Option Behavior

--branch [branch_name] option:

  • Adds a branch creation task at the beginning of the task list
  • If branch_name is provided: Uses the specified branch name
  • If branch_name is omitted: Auto-generates branch name following Git naming conventions
  • All commits during task execution will be made to this branch
  • Can be used independently without --pr
  • Useful when you want to work on a feature branch but don't need a PR yet

Branch name auto-generation rules:

  • Analyzes TODO file content to determine branch type and purpose
  • Follows Git naming conventions: {type}/{descriptive-name}
  • Types: feature/, bugfix/, refactor/, chore/, docs/
  • Format: lowercase, hyphen-separated, English
  • Example: feature/actionlog-email-notification

--pr option:

  • Includes all --branch functionality (branch creation and commits)
  • Adds a pull request creation task at the end of the task list
  • If --branch is not specified, a branch name will be auto-generated
  • The PR will include all changes made during task execution
  • If PR template instructions exist in CLAUDE.md or similar files, those templates will be used

Usage Examples

# Example 1: Create branch with auto-generated name (no PR)
/todo-task-planning TODO.md --branch
# → Generates branch name like: feature/actionlog-notification

# Example 2: Create branch with specific name (no PR)
/todo-task-planning docs/todos/feature-x.md --branch feature/user-auth

# Example 3: Create PR with auto-generated branch name
/todo-task-planning docs/todos/feature-x.md --pr
# → Auto-generates branch name and creates PR

# Example 4: Create PR with specific branch name
/todo-task-planning docs/todos/feature-x.md --pr --branch feature/user-auth

# Example 5: Basic task planning (no branch, no PR)
/todo-task-planning docs/todos/feature-x.md

Implementation Guidance

When these options are specified, the task planning should include:

For --branch option:

  • Branch Name Determination:

- If branch name is provided: Use as-is (validate against naming conventions) - If branch name is omitted: Auto-generate following this logic: 1. Read TODO file title and content 2. Determine branch type based on task nature: - feature/ - New functionality implementation - bugfix/ - Bug fixes, issue resolution - refactor/ - Code restructuring without behavior change - chore/ - Development environment, dependencies, tooling - docs/ - Documentation updates 3. Extract key feature/issue name from TODO (2-4 words max) 4. Convert to lowercase, hyphen-separated English 5. Format: {type}/{descriptive-name} 6. Example: "ActionLog Email Notification" → feature/actionlog-email-notification

  • Task to create the determined/generated branch at the beginning
  • All modification tasks should indicate they will be committed to this branch
  • No PR-related tasks

For --pr option:

  • Task to create a branch (using specified name or auto-generated)
  • All modification tasks with commit instructions
  • Final task to create a pull request with proper description
  • PR description should summarize all changes made

Reference Documentation

Core Guidelines

Phase 0 reads and applies /key-guidelines at the start of every execution

[CRITICAL]Important Implementation Requirements

MANDATORY: This command MUST update the $ARGUMENTS file (the file specified as a parameter)

  • Main Claude executor (not a subagent) uses Edit or Write tool to update files
  • After calling subagents in Phase 1, update the $ARGUMENTS file with those results in Phase 5
  • Add new task planning results in a structured format while preserving existing content
  • After file update is complete, confirm, verify, and report the updated content
  • CRITICAL: The $ARGUMENTS file update is NOT optional - it must be executed in every run

Processing Flow

[CRITICAL]Phase Dependency and Execution Rules

[PROHIBITED]PROHIBITED PHASE SHORTCUTS - MUST FOLLOW SEQUENTIAL FLOW ⛔

The following phase shortcuts are STRICTLY PROHIBITED:

  • [NG]Phase 1 → Phase 5 direct transition (MOST COMMON VIOLATION)
  • [NG]Phase 1 → Phase 6 direct transition
  • [NG]Phase 2 → Phase 5 direct transition (skipping Phase 3-4)
  • [NG]Phase 3 → Phase 5 direct transition (skipping Phase 4)

Why These Shortcuts Are Dangerous:

  • Data Loss: Phase 2 existing task analysis will be ignored
  • No Integration: New tasks won't integrate with existing TODO.md content
  • Missing Validation: User questions won't be asked, leading to incorrect assumptions
  • File Creation Failures: docs/memory files won't be created at proper time

📊 Phase Dependency Flow Chart

┌─────────────────────────────────────────────────────────────────┐
│                    PHASE EXECUTION FLOW                         │
│                 (MUST FOLLOW SEQUENTIAL ORDER)                  │
└─────────────────────────────────────────────────────────────────┘

Phase 0: Key Guidelines Reading → [詳細](PHASE-0-KEY-GUIDELINES.md)    [OK]MANDATORY
└─ Read and apply key-guidelines skill
   └─ Output: guidelines_loaded
              ↓
Phase 1: TODO File Reading → [詳細](PHASE-1-TODO-READING.md)          [OK]MANDATORY
└─ Read and parse $ARGUMENTS file
   └─ Parse --pr and --branch options
   └─ Output: HAS_PR_OPTION, HAS_BRANCH_OPTION, BRANCH_NAME
              ↓
Phase 2: Explore Subagent → [詳細](PHASE-2-EXPLORE.md)                [OK]MANDATORY
└─ Call Task tool (Explore)
   └─ Output: exploration_results
              ↓
Phase 3: Plan Subagent → [詳細](PHASE-3-PLAN.md)                      [OK]MANDATORY
└─ Call Task tool (Plan) with exploration_results
   └─ Output: planning_results
              ↓
Phase 4: project-manager skill → [詳細](PHASE-4-PROJECT-MANAGER.md)   [OK]MANDATORY
└─ Call Skill tool (project-manager) with exploration + planning results
   └─ Output: strategic_plan
              ↓
Phase 5: Verification → [詳細](PHASE-5-VERIFICATION.md)               [OK]MANDATORY
└─ Verify all subagents completed successfully
   └─ Verify: exploration_results, planning_results, strategic_plan exist
              ↓
Phase 6: File Analysis → [詳細](PHASE-6-ANALYSIS.md)                  [OK]MANDATORY
└─ Read $ARGUMENTS file
   └─ Output: existingTasks, taskProgress
              ↓
Phase 7: Task Analysis & Breakdown → [詳細](PHASE-7-BREAKDOWN.md)     [OK]MANDATORY
└─ Input: Phase 1-5 results + Phase 6 results
   └─ Output: Task breakdown, feasibility analysis
              ↓
Phase 8: Question Management → [詳細](PHASE-8-QUESTIONS.md)           [WARNING]CONDITIONAL
├─ CONDITION A: Questions exist         [OK]MANDATORY
│  └─ Execute AskUserQuestion tool
│  └─ Wait for user responses
│  └─ Create questions.md file
│  └─ Output: User decisions recorded
├─ CONDITION B: No questions            [OK]ALLOWED (Must document reason)
│  └─ Proceed to Phase 9
│  └─ Document why no questions needed
└─ [PROHIBITED]GATE: Phase 9 entrance checkpoint
              ↓
Phase 9: File Update → [詳細](PHASE-9-UPDATE.md)                      [OK]MANDATORY
├─ Create docs/memory files (exploration, planning, questions)
├─ Update $ARGUMENTS file with task checklist
├─ Conditional branch/PR task insertion:
│  └─ Evaluate HAS_BRANCH_OPTION and HAS_PR_OPTION
│  └─ Insert branch creation task if needed (beginning)
│  └─ Insert PR creation task if needed (end)
└─ Output: Complete TODO.md with all tasks
              ↓
Phase 10: Verification & Feedback → [詳細](PHASE-10-VERIFICATION.md)  [OK]MANDATORY
└─ Verify file updates, AskUserQuestion execution
   └─ Report to user

🔐 Phase Requirement Markers

PhaseStatusSkippable?DependenciesCritical Output
Phase 0[OK]MANDATORY🚫 NONoneguidelines_loaded
Phase 1[OK]MANDATORY🚫 NOPhase 0HAS_PR_OPTION, HAS_BRANCH_OPTION, BRANCH_NAME
Phase 2[OK]MANDATORY🚫 NOPhase 1exploration_results
Phase 3[OK]MANDATORY🚫 NOPhase 2planning_results
Phase 4[OK]MANDATORY🚫 NOPhase 3strategic_plan
Phase 5[OK]MANDATORY🚫 NOPhase 4Verification status
Phase 6[OK]MANDATORY🚫 NOPhase 5existingTasks, taskProgress
Phase 7[OK]MANDATORY🚫 NOPhase 1-6Task breakdown, feasibility
Phase 8[WARNING]CONDITIONAL🚫 NO (See conditions)Phase 7User decisions (if questions exist)
Phase 9[OK]MANDATORY🚫 NOPhase 1-8Updated $ARGUMENTS file, docs/memory files, branch/PR tasks inserted
Phase 10[OK]MANDATORY🚫 NOPhase 9Verification report

Phase 8 Conditions:

  • [OK]Questions exist: MUST execute AskUserQuestion tool and wait for responses
  • [OK]No questions: MUST proceed to Phase 9 and document reason in Phase 10

[WARNING]Critical Phase Transition Rules

Rule 1: Sequential Execution Only

  • Each phase MUST complete before the next phase begins
  • No parallel execution of phases
  • No skipping of phases
  • Each Phase is independent and must be executed in a separate turn/message

Rule 2: Subagent Phases (2-4) Must Execute Sequentially

[NG]WRONG FLOW (Parallel Execution):
Phase 2 (Explore) + Phase 3 (Plan) + Phase 4 (project-manager) → Called in same message
                    ↓
            PARALLEL EXECUTION
                    ↓
    Result: Phase 3/4 start before Phase 2 completes, missing dependencies

[OK]CORRECT FLOW (Sequential Execution):
Phase 0 → Phase 1 → Phase 2 (wait) → Phase 3 (wait) → Phase 4 (wait) → Phase 5 → ...
   ↓         ↓         ↓                ↓                ↓                ↓
  Key-    TODO     Explore          Plan          project-mgr      Verify
 Guide   Reading   (output)       (needs P2)      (needs P2+P3)   (check all)

Rule 3: Never Skip to Phase 9 (File Update)

  • Phase 9 requires outputs from ALL previous phases
  • Skipping Phase 6-8 causes data loss and missing integration
  • Phase 6 (File Analysis) results are mandatory for Phase 7
  • Phase 8 (Questions) is the entrance gate to Phase 9

Rule 4: Phase 8 is a Mandatory Checkpoint

  • Even if no questions exist, Phase 8 MUST be executed to document this fact
  • Phase 9 entrance gate verifies Phase 8 completion

Variable Scope and Persistence

IMPORTANT: Variables set in Phase 1 persist throughout all subsequent phases (Phase 2-10).

Phase 1 Variables Used in Later Phases:

  • HAS_PR_OPTION, HAS_BRANCH_OPTION, BRANCH_NAME, IS_AUTO_GENERATED

- Set in: Phase 1 Step 2 (argument parsing) - Used in: Phase 9 (conditional task insertion) - Scope: Available throughout entire skill execution

Variable Lifecycle:

Phase 1 → Set variables
    ↓
Phase 9 → Use variables for conditional logic

Related Documentation

For detailed information about each phase, see:

Additional resources:

  • Advanced Usage - Iterative Execution, Best Practices, Common Mistakes
  • Examples - Usage Examples and Output Formats

平台分布

Codex

35.05%
按下载量换算60

Claude

31.08%
按下载量换算53

Cursor

19.71%
按下载量换算34

Gemini CLI

8.01%
按下载量换算14

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。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills