markdown-task-manager

name	markdown-task-manager
description	Use when managing tasks, the system is a Kanban task manager based on local Markdown files (`kanban.md` and `archive.md`). It follows a strict format compatible with the task-manager.html web application.

📋 Markdown Task Manager Skill

When to Use This Skill

• Create tasks to document work to be done
• Plan complex features
• Track project progress
• Archive completed tasks
• Generate status reports

📋 STRICT Task Format

Mandatory Template

### TASK-XXX | Task title

**Status**: <column-id>
**Priority**: [Critical|High|Medium|Low] | **Category**: [Value] | **Assigned**: @user1, @user2
**Created**: YYYY-MM-DD | **Started**: YYYY-MM-DD | **Due**: YYYY-MM-DD | **Finished**: YYYY-MM-DD
**Tags**: #tag1 #tag2 #tag3

Free text description. **NO `##` or `###` headings allowed**.

**Subtasks**:
- [ ] First subtask
- [x] Completed subtask

**Notes**:
Additional notes with subsections `**Title**:`.

**Result**:
What was done.

**Modified files**:
- file.js (lines 42-58)

❌ FORBIDDEN

• ## Title or ### Title inside a task
• **Subtasks** or **Notes** without :
• Automatic archiving (only on user request)

Why? All tasks share one ## Tasks section, so a stray ##/### heading inside a task can break parsing.

🔄 Workflow

V2 — status is a field, not a position. A task's column is its **Status**: value; all tasks live in one ## Tasks section. Move a task by editing its one **Status**: line — never cut/paste the whole block (~1 line vs 20+).

1. New Request

1. Read kanban.md to get the last task ID
2. Add the task to the ## Tasks section with **Status**: todo (first column id)
3. Unique ID (TASK-XXX) auto-incremented
4. Break down into subtasks if needed
5. Increment counter in <!-- Config: Last Task ID: XXX -->

2. Start Work

1. Edit one line: **Status**: todo → **Status**: in-progress
2. Add **Started**: YYYY-MM-DD
3. Check off subtasks progressively

3. Finish Work

1. Edit one line: **Status**: in-progress → **Status**: done
2. Add **Finished**: YYYY-MM-DD
3. Document in **Notes**::
   • **Result**: - What was done
   • **Modified files**: - List with line numbers
   • **Technical decisions**: - Choices made
   • **Tests performed**: - Validated tests

4. Archiving

⚠️ Tasks are NOT archived immediately!

• Completed tasks stay at **Status**: done
• Only on user request → move to archive.md
• Never archive directly at end of work

📁 File Structure

kanban.md

# Kanban Board

<!-- Config: Last Task ID: 42 -->
<!-- Format: v2 -->

## ⚙️ Configuration

**Columns**: 📝 To Do (todo) | 🚀 In Progress (in-progress) | 👀 Review (in-review) | ✅ Done (done)
**Categories**: Frontend, Backend, DevOps
**Users**: @alice, @bob
**Priorities**: 🔴 Critical | 🟠 High | 🟡 Medium | 🟢 Low
**Tags**: #bug, #feature, #docs

---

## Tasks

### TASK-001 | Title
**Status**: todo
[...]

### TASK-003 | Completed task
**Status**: done
[...]

All tasks share the single ## Tasks section; the column comes from **Status**: (which must equal a column id from the Config), and file order = order within a column.

⚡ Token-economical operations (for AI)

No stored index — rebuild it on demand with grep, and read/edit one task:

grep -nE "^### TASK-|^\*\*Status\*\*:" kanban.md    # board overview (index on demand)
grep -n "^### TASK-042 " kanban.md                   # locate a task by id
sed -n '/^### TASK-042 /,/^### TASK-/p' kanban.md     # read just that task
# MOVE = edit one line:   **Status**: todo  ->  **Status**: in-progress
grep -n "^\*\*Status\*\*: in-progress" kanban.md | head -1   # next task in a column

archive.md

# Task Archive

> Archived tasks

## ✅ Archives

### TASK-001 | Archived task
[... full content ...]

---

### TASK-002 | Another archived task
[... full content ...]

🎯 Golden Rules

✅ ALWAYS

1. Create task BEFORE coding
2. Strict format (no ## inside tasks)
3. Break down if complex (3+ steps)
4. Update in real-time
5. Document result in **Notes**:
6. Reference tasks in commits (TASK-XXX)
7. Leave in "Done" (archive only on user request)

❌ NEVER

1. ## Title inside a task
2. Code without creating task
3. Forget to check off subtasks
4. Archive immediately (stay in "Done")
5. Forget to document the result

🔧 User Commands

Planning

• "Plan [feature]"
• "Create roadmap for 3 months"

Execution

• "Do TASK-XXX"
• "Continue TASK-XXX"

Tracking

• "Where are we?"
• "Weekly status"

Modifications

• "Break down TASK-XXX"
• "Add subtask to TASK-XXX"

Search

• "Search in archives: [keyword]"

Maintenance

• "Archive completed tasks"

📝 Complete Examples

Simple Task

### TASK-001 | Fix login bug

**Priority**: Critical | **Category**: Backend | **Assigned**: @bob
**Created**: 2025-01-20 | **Due**: 2025-01-21
**Tags**: #bug #urgent

Users cannot log in. Error 500 in logs.

**Notes**:
Check Redis, related to yesterday's deployment.

Complete Task with Result

### TASK-042 | Notification system

**Priority**: High | **Category**: Backend | **Assigned**: @alice
**Created**: 2025-01-15 | **Started**: 2025-01-18 | **Finished**: 2025-01-22
**Tags**: #feature

Real-time notifications with WebSockets.

**Subtasks**:
- [x] Setup WebSocket server
- [x] REST API
- [x] Email sending
- [x] Notifications UI
- [x] E2E tests

**Notes**:

**Result**:
✅ Functional system with WebSocket, REST API and emails.

**Modified files**:
- src/websocket/server.js (lines 1-150)
- src/api/notifications.js (lines 20-85)

**Technical decisions**:
- Socket.io for WebSockets
- SendGrid for emails
- 30-day history in MongoDB

**Tests performed**:
- ✅ 100 simultaneous connections
- ✅ Auto-reconnection
- ✅ Emails < 2s

🛠️ Skill Functions

When using this skill, you must:

1. Read kanban.md to understand current state and get last ID
2. Create tasks following strict format
3. Update tasks by moving between sections
4. Check off subtasks progressively
5. Document result in Notes before marking Done
6. Increment Last Task ID in config comment
7. Never archive unless explicitly requested

📘 Git Integration

# Commits with reference
git commit -m "feat: Add feature (TASK-042 - 3/5)"
git commit -m "fix: Bug fix (TASK-001)"

# Branches
git checkout -b feature/TASK-042-notifications

⚠️ Critical Points of Attention

1. Markdown Format: Strictly respect format (no ## inside tasks)
2. ID Increment: Always increment <!-- Config: Last Task ID: XXX -->
3. Columns: Use exact column names defined in Configuration
4. Archiving: NEVER archive automatically, only on request
5. Documentation: Always fill **Notes**: with Result, Modified files, etc.

🎓 Usage

Initialization

Before using the skill, verify the project contains:

• kanban.md (required)
• archive.md (required)
• AI_WORKFLOW.md (optional but recommended)

First Use

"Use the markdown-task-manager skill to create a task for [feature]"

Invocation Examples

"Skill markdown-task-manager: create a task to implement authentication"
"Skill markdown-task-manager: update TASK-007 with results"
"Skill markdown-task-manager: list all tasks in progress"
"Skill markdown-task-manager: archive completed tasks"

🔍 Implementation Details

Reading kanban.md

Always start by reading kanban.md to:

• Get the last task ID from <!-- Config: Last Task ID: XXX -->
• Understand existing columns structure
• Check current tasks in each column

Creating a New Task

1. Calculate new ID: last_id + 1
2. Format as TASK-XXX (3 digits with leading zeros)
3. Add to "📝 To Do" section
4. Update <!-- Config: Last Task ID: XXX --> comment
5. Use today's date for **Created**:

Moving Tasks Between Columns

When moving a task:

1. Copy entire task content (from ### TASK-XXX to blank line before next task)
2. Paste in target column section
3. Delete from original location
4. Update dates accordingly (**Started**: or **Finished**:)

Completing Tasks

Before moving to "✅ Done":

1. Ensure all subtasks are checked [x]
2. Add **Finished**: YYYY-MM-DD
3. Fill in **Notes**: section with:
   • **Result**: describing what was accomplished
   • **Modified files**: listing changed files with line ranges
   • **Technical decisions**: if any choices were made
   • **Tests performed**: if tests were run

Archiving Tasks

Only when user explicitly requests archiving:

1. Read task from "✅ Done" section in kanban.md
2. Append task to "## ✅ Archives" section in archive.md
3. Add separator --- between archived tasks
4. Remove task from kanban.md

---

This skill ensures complete traceability and total transparency of work done by AI.