← 返回 Skills 市场
yayayahei

Eisenhower Task Manager

作者 yayayahei · GitHub ↗ · v1.0.9 · MIT-0
cross-platform ⚠ suspicious
331
总下载
0
收藏
1
当前安装
10
版本数
在 OpenClaw 中安装
/install eisenhower-task-manager
功能描述
Task management based on Eisenhower Matrix + P0-P2 priority with Customer Project Management. Four quadrants for execution, separate Customer Project List fo...
使用说明 (SKILL.md)

Task Manager - Four-Level System

Architecture

Layer File Purpose
Customer tasks/customer-projects.md ALL customer projects (single source of truth)
Execution tasks/tasks.md Four quadrants (personal + customer tasks requiring YOUR execution)
Future tasks/maybe.md Ideas for later evaluation
Delegation tasks/delegation.md Tasks assigned to others
Archive tasks/archived.md Completed/deleted history

Core Principles

  1. Customer projects NEVER duplicate - exist only in Customer Project List
  2. Personal execution only - customer tasks enter quadrants only when YOU must execute
  3. Sequential numbering - 1-N per file, no gaps, no duplicates
  4. Auto-fix numbering - after EVERY operation, verify and fix sequence

Four Quadrants

Q Name Strategy When
Q1 Important + Urgent Execute immediately Deadlines, blockers
Q2 Important + Not Urgent Plan & schedule OKRs, strategic work
Q3 Not Important + Urgent Batch process Maintenance, delegatable
Q4 Not Important + Not Urgent Postpone/delete Exploration, optional

Priority Tiers (within quadrant)

  • P0: Today (hard deadline or blocking)
  • P1: This week
  • P2: This month/quarter

Customer Project Workflow

New Customer Project
    ↓
Add to Customer Project List (record all)
    ↓
Need YOUR execution?
    ├─ YES → Add to appropriate Quadrant, tag with [Customer/ProjectName]
    └─ NO → Keep in Customer Project List only
    ↓
Regular Review → Update status (Active/Blocked/Pending)
    ↓
Completed → Archive entry + complete quadrant task if exists

Task Numbering Rules

Quadrant Tasks (tasks.md)

  • Q1: 1 to N
  • Q2: (N+1) to M
  • Q3: (M+1) to P
  • Q4: (P+1) to Q
  • Continuous sequence across quadrants, no gaps

Maybe List (maybe.md)

  • Sequential: 1, 2, 3...
  • ALWAYS append to END - never insert in middle
  • Verify: last_number == total_count

Customer Projects (customer-projects.md)

  • Per-customer numbering: 1, 2, 3...
  • Independent across customers

Delegation List (delegation.md)

  • Sequential: 1, 2, 3...
  • ALWAYS append to END
  • Verify: last_number == total_count
  • Format: ### X. Task Name @assignee

Archive (archived.md)

  • Sequential: 1, 2, 3...
  • ALWAYS append to END
  • Group by archive date
  • Format: ### X. Task Name [Archived: YYYY-MM-DD]

Numbering Verification (MUST RUN AFTER EVERY OPERATION)

  1. Read target file
  2. Scan for: duplicates, gaps, wrong transitions
  3. Reassign sequential numbers starting from 1
  4. Re-read to verify
  5. Repeat if issues remain

Numbering Checklist:

  • Sequential from 1, no gaps
  • No duplicates
  • Q2 starts at Q1_count+1, Q3 at Q2_end+1, etc.

Operations

Add Task

See references/task-add.md for complete workflow.

Quick Guide:

  1. Determine level: Quadrants / Delegation / Maybe
  2. Determine quadrant (if applicable): Q1/Q2/Q3/Q4
  3. Assign priority: P0/P1/P2
  4. Insert at appropriate position
  5. Renumber ALL tasks (sequential from 1)
  6. Update statistics
  7. Display updated list
  8. (Optional) Offer to open Dashboard for visual review

Complete Task

See references/task-complete.md for complete workflow.

Key Points:

  • MUST NOT stop at just marking [x] - full flow required
  • Delete completed task from source file
  • Renumber remaining tasks
  • Move to archived.md
  • Update statistics
  • Display updated list
  • (Optional) Offer to open Dashboard to see updated status

Three Scenarios:

  1. Main Task List (tasks.md) - Four Quadrant tasks
  2. Delegation List (delegation.md) - Delegated tasks
  3. Maybe List (maybe.md) - Future tasks

Post-Operation Dashboard Offer

After any task operation (add/complete/update), the skill SHOULD offer to open the dashboard:

✅ Task Added Successfully

**Summary:**
- Task: [Task Name]
- Location: Q2 #15
- Priority: P1

📊 Would you like to view the task dashboard?
→ User approves: Launch dashboard and open browser
→ User declines: Continue with next operation

Implementation: Check if user wants to open dashboard after major operations.

Add Customer Project

Step 1: Add to Customer Project List

### [Customer Name]

#### X. Project Name [Status]
- **Status**: Active / Blocked / Pending / Completed
- **Type**: Implementation / Support / POC / Maintenance
- **Description**: Brief
- **Created**: YYYY-MM-DD
- **Last Review**: YYYY-MM-DD
- **Notes**: Progress, blockers

Step 2: If personal execution needed, also add to Quadrants

Complete Customer Project

  1. Mark Completed in Customer Project List
  2. Archive the entry in archived.md
  3. Complete related quadrant task if exists

Add to Maybe List

  1. Read tasks/maybe.md
  2. Find highest task number (N)
  3. Append new task as #(N+1) at END
  4. Update total count and timestamp

Check/Fix Numbering

Trigger phrases: "check task numbering", "fix sequence", "序号问题"

Action: Read file → scan issues → auto-fix → re-verify → repeat if needed

Output Format (MUST OUTPUT AFTER EVERY OPERATION)

✅ [Operation] Completed

**Summary:**
- Operation: [Add/Update/Complete/Move/Delete]
- Target: [Task name/ID]
- Location: [Q1/Q2/Q3/Q4/Customer/Maybe/Delegation/Archive]

**Statistics:**
| Quadrant | Count |
|----------|-------|
| Q1 | X |
| Q2 | X |
| Q3 | X |
| Q4 | X |

See references/output-examples.md for detailed examples.

Healthy Metrics

List Target Warning
Customer Projects 5-15 -
Q1 3-5 >5 indicates poor planning
Q2 8-12 Core investment area
Q3 5-10 Automate gradually
Q4 \x3C10 Clean up regularly
Maybe \x3C20 Review monthly
Delegation \x3C15 Review weekly

Time Rhythm

  • Daily: Check Q1, customer escalations
  • Weekly: Review Customer Projects, Q2 planning, Delegation check
  • Monthly: Archive completed, review blocked, adjust priorities

Task Tags

  • Customer tasks in quadrants: [Customer/ProjectName]
  • Blocked tasks: 🚫 Blocked
  • Status: P0/P1/P2

Dashboard (Optional)

A real-time web dashboard is available to view all tasks in a friendly format.

Location: dashboard/

Features:

  • 📊 Real-time statistics showing task counts by quadrant
  • 🔥 Four-quadrant matrix view with color coding
  • 🏢 Customer projects overview with filtering
  • 👑 Delegation list with assignee and deadline info
  • 🌱 Maybe list for future ideas
  • 🎨 Dark theme with modern UI design
  • 🌐 Bilingual support (English/Chinese)
  • ⚡ Real-time updates via WebSocket when markdown files change
  • 📱 Responsive layout for different screen sizes
  • 💡 Hover to see full task details - Mouse over any task card to see complete information including description, subtasks, tags, and timestamps

Dashboard Launch Workflow

After completing task operations, the skill will offer to open the dashboard:

✅ [Operation] Completed

Would you like to view the task dashboard?
[Open Dashboard] [Skip]

If approved → Auto-start dashboard and open browser
If skipped  → Dashboard remains available for manual launch

Manual Launch (Anytime)

Method 1: Daemon Mode (Recommended for Long-term Use)

Run in background, service continues after terminal closes:

cd dashboard
./start.sh --daemon     # Start in background, no terminal dependency

Features:

  • ✅ Service continues after terminal closes
  • ✅ Auto PID management, prevents duplicate starts
  • ✅ Logs output to dashboard.log
  • ✅ Shows access URL and PID after startup

Stop the daemon:

./stop.sh               # Gracefully stop the service

Method 2: Foreground Mode (For Debugging)

Suitable for development and debugging, service stops when terminal closes:

cd dashboard
./start.sh              # Run in foreground, Ctrl+C to stop

Advanced Usage

# Specify port (default 8080)
./start.sh --daemon --port 3000

# Check running status
curl http://localhost:8080/api/health

Architecture

  • Markdown files are the single source of truth
  • Node.js server parses markdown in real-time
  • WebSocket pushes updates to browser when files change
  • No database required - pure markdown-driven
  • Enhanced error handling: global exception capture, port conflict detection

Reference Materials

File Purpose
references/task-add.md Complete workflow for adding tasks - levels, quadrants, priorities, renumbering
references/task-complete.md Complete workflow for completing tasks - 3 scenarios (Main/Delegation/Maybe), archiving
references/numbering-rules.md Numbering rules for all 5 task lists
references/output-examples.md Output format examples
references/maybe-list-workflow.md Step-by-step Maybe List operations
references/templates.md Statistics and report templates
references/dashboard-offer.md Dashboard offer workflow after task operations
dashboard/ Real-time web dashboard (./start.sh --daemon for background mode)

Critical: Always read task-add.md or task-complete.md before performing those operations to ensure full workflow compliance.


Version: 8.2 (Added daemon mode, stop.sh, enhanced error handling) Last Updated: 2026-04-21

安全使用建议
Things to check before installing or running this skill: - Review server.js and start.sh: confirm the server listens on localhost (127.0.0.1) rather than 0.0.0.0, and verify there are no unexpected external network calls or endpoints. - Confirm you have Node.js/npm and are comfortable the start script will run 'npm install' in the skill directory (it will download packages from the public npm registry). - Inspect the code that reads/writes files (parser.js and whatever performs renumber/delete/archive operations) to ensure it only touches the intended tasks/*.md files and does not access other paths. - Be aware the skill will modify your markdown files (delete/move/renumber) as part of 'Complete Task' and 'Auto-fix numbering' flows — back up your tasks/ directory before first run. - If you cannot review the server code, run it in a sandboxed environment or container to limit filesystem and network exposure. - The registry metadata should declare required binaries (node/npm) and an install spec; the absence of these is an inconsistency to flag with the skill author.
功能分析
Type: OpenClaw Skill Name: eisenhower-task-manager Version: 1.0.9 The Eisenhower Task Manager skill is a legitimate tool for managing tasks across multiple markdown files with a supporting Node.js web dashboard. The bundle includes shell scripts (start.sh, stop.sh) and a server (server.js) using Express and WebSockets to provide real-time visual updates of task quadrants. While the skill requires the ability to execute local shell commands and run a background process for the dashboard, these actions are well-documented, aligned with the stated functionality, and include basic security practices such as HTML escaping in the frontend (app.js) to prevent XSS.
能力评估
Purpose & Capability
The skill claims to be instruction-only with no required binaries/config, but the bundle contains a Node.js dashboard (server.js, parser.js, start.sh, check-and-install.sh, package.json). Running the dashboard requires Node/npm and will read/write files under a tasks/ directory — reasonable for a task manager but inconsistent with the declared 'no required binaries' and 'instruction-only' metadata.
Instruction Scope
SKILL.md instructs the agent to read, modify, renumber, delete, and archive markdown task files (tasks.md, customer-projects.md, maybe.md, delegation.md, archived.md) after every operation. Those are destructive file operations (delete/move/renumber). It also directs offering to 'launch dashboard and open browser' which implies starting a local server and opening a browser. The file modifications and auto-launch behavior are within the skill's purpose but are high-impact and must be explicitly consented to.
Install Mechanism
There is no formal install spec in the registry metadata, but the package includes start/check-and-install scripts that run 'npm install' and depend on common npm packages (express, ws, chokidar). Installing pulls from the public npm registry (package-lock.json shows registry URLs). This is a normal flow for a Node app but the registry entry should have declared this dependency and required binaries (node/npm).
Credentials
The skill does not request environment variables, credentials, or external endpoints. Code reads local markdown files from a tasks/ directory and uses local npm packages — this is proportionate to its stated purpose.
Persistence & Privilege
always:false (normal) and no declared credential persistence. However, the dashboard starts a local HTTP/WebSocket server that will serve the parsed task data and watch files for changes (chokidar). You should confirm the server binds only to localhost and does not expose files to the network unintentionally.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install eisenhower-task-manager
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /eisenhower-task-manager 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.9
Translate SKILL.md to English
v1.0.8
Add daemon mode for dashboard, stop.sh script, and enhanced error handling
v1.0.7
**Dashboard view improvement with hover for full task details.** - Added mouse hover support in the dashboard to display complete task information (description, subtasks, tags, timestamps). - No functional changes to core task management logic. - SKILL.md documentation updated to describe the new dashboard hover feature.
v1.0.6
No user-facing or functional changes in this version. - No changes detected in any files. - Behavior and documentation remain unchanged from previous version.
v1.0.5
No functional or documentation changes in this version. - No file changes detected compared to previous release. - Existing functionality and documentation remain unchanged.
v1.0.4
**Major update: Dashboard integration for real-time visual task management.** - Added a new `dashboard/` web app for live task overview, including statistics, quadrant views, and project lists. - Implemented post-task-operation prompts to offer users the option to open the dashboard for visual review. - Dashboard features: real-time WebSocket updates, bilingual support (EN/zh), dark theme, mobile responsive, auto-installation on first run. - No backend database required—dashboard reads and visualizes markdown files directly. - Core task management and numbering workflows unchanged; dashboard is fully optional.
v1.0.3
**Concise Changelog for eisenhower-task-manager v1.0.3** - Major specification refactor: all task management architecture, rules, and workflows are now clarified and condensed for easier reference. - Added dedicated reference files: numbering rules, output formats, maybe-list workflow, and templates for reporting/statistics. - Detailed, strict task numbering and verification rules introduced; must auto-correct after every operation. - Output requirements are formalized with clear summary and statistics templates. - All core operation flows (add, complete, move, archive tasks/projects) now have linked step-by-step documentation. - Emphasis on reducing duplication, enforcing single source of truth for customer projects, and maintaining healthy task distribution.
v1.0.2
**Customer project management integrated as a dedicated layer.** - Added a separate Customer Project List to manage all customer projects outside the four quadrants. - Clear workflow for customer tasks: only add to quadrants if personal execution is required; all others remain in the Customer Project List. - Updated task file structure to include `tasks/customer-projects.md`. - Detailed process for reviewing, tracking, and archiving customer projects. - Enhanced rules for task numbering: automatic verification and fixing after each task operation. - Updated usage and operation flows to reflect the new customer management process.
v1.0.1
No changes detected in this version. - Version bumped to 1.0.1 for administrative or metadata reasons. - No feature, documentation, or functionality updates present.
v1.0.0
Major overhaul: Fully redesigned task management system from previous SQLite-based implementation to an Eisenhower Matrix + three-level priority model. - Shifted from code/database focus to a Markdown-driven, principle-first approach for personal and team task management. - Introduced a three-level structure: Execution Layer (Four Quadrants), Future Layer (Maybe List), and Management Layer (Delegation List). - Detailed scenarios and guidelines for when and how to add tasks to each level. - Added clear priority sorting (P0/P1/P2) within quadrants, and explicit principles for review, adjustment, and task flow. - Task file structure now explained via markdown grouping and rotation mechanisms; removed code and JSON/meta requirements.
元数据
Slug eisenhower-task-manager
版本 1.0.9
许可证 MIT-0
累计安装 1
当前安装数 1
历史版本数 10
常见问题

Eisenhower Task Manager 是什么?

Task management based on Eisenhower Matrix + P0-P2 priority with Customer Project Management. Four quadrants for execution, separate Customer Project List fo... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 331 次。

如何安装 Eisenhower Task Manager?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install eisenhower-task-manager」即可一键安装,无需额外配置。

Eisenhower Task Manager 是免费的吗?

是的,Eisenhower Task Manager 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Eisenhower Task Manager 支持哪些平台?

Eisenhower Task Manager 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 Eisenhower Task Manager?

由 yayayahei(@yayayahei)开发并维护,当前版本 v1.0.9。

💬 留言讨论