tasks = [
    {
        "id": "research-a",
        "name": "调研方案A",
        "complexity": 4.2,
        "dependencies": [],
        "status": "pending"
    },
    {
        "id": "summary",
        "name": "汇总分析",
        "complexity": 5.0,
        "dependencies": ["research-a"],
        "status": "pending"
    }
]

for task in tasks:
    workflow.add_task(task)

3. Analyze and Schedule

# Build dependency DAG and schedule batches
schedule = workflow.analyze_and_schedule()

# Output:
# Batch 1: research-a (complexity: 4.2)
# Batch 2: summary (complexity: 5.0, depends on: research-a)

4. Update Progress

workflow.update_task_status("research-a", "completed")
workflow.update_task_status("summary", "running")

5. Check Status

status = workflow.get_status()
print(f"Completed: {status['completed']}/{status['total']}")

When to Use

Use this skill when:

Managing projects with 5+ interdependent tasks
Need automatic dependency resolution
Want complexity-based task ordering
Require persistent progress tracking
Working across multiple sessions

Skip when:

Simple sequential tasks without dependencies
Quick one-off tasks
Single-session work

File Format

Workflow stores data in daily markdown files:

~/.openclaw/workspace/task_backlog/
├── task-workflow-progress-2026-02-18.md  # Yesterday (archived)
├── task-workflow-progress-2026-02-19.md  # Today (active)
└── task-workflow-progress-2026-02-20.md  # Tomorrow (pre-created)

See references/file-format.md for complete format specification.

Key Features

Feature	Description
DAG Scheduling	Topological sort for dependency resolution
Complexity Scoring	1-10 scale, lower = batch earlier
Dynamic Insertion	Add tasks mid-execution without restart
Auto Migration	CST 0:00 moves pending tasks to new day
Cron Integration	OpenClaw Cron API support

Dependencies

Task dependencies use simple ID references:

{
    "id": "summary",
    "dependencies": ["research-a", "research-b"]
}

The system builds a DAG and schedules tasks in optimal batches.

Complexity Scoring

Score tasks 1-10 based on:

1-3: Simple, well-understood tasks
4-6: Moderate complexity, some unknowns
7-10: Complex, high uncertainty

Lower complexity tasks are batched first.

Architecture

TaskWorkflowV3
├── init_daily_file()      # Create daily markdown
├── add_task()             # Add to queue
├── analyze_and_schedule() # Build DAG, create batches
├── update_task_status()   # Update progress
├── get_status()           # Report statistics
└── migrate_tasks()        # Daily auto-migration

Cron Configuration

# config/cron.yaml
cron_jobs:
  - name: task-workflow-daily-init
    schedule: "0 0 * * *"  # CST 0:00
    command: "python -m task_workflow_v3.cli init-daily"
  
  - name: task-workflow-cleanup
    schedule: "0 1 * * *"
    command: "python -m task_workflow_v3.cli cleanup --days 30"

V2 to V3 Migration

Feature	V2	V3
File persistence	❌	✅ Markdown files
Progress tracking	❌ Session-only	✅ File-based
Auto archive	❌	✅ Daily migration
Cron integration	❌	✅ OpenClaw API

See references/v3-migration.md for full migration guide.

References

File Format: references/file-format.md
Migration Guide: references/v3-migration.md
API Documentation: See Python docstrings

Task Workflow V3 — Intelligent scheduling for complex projects.

task-workflow

Task Workflow V3

Quick Start

1. Initialize Workflow

2. Add Tasks