task-based-multiagent

# Task-Based Multi-Agent Skill

Guide creation of task-based multi-agent systems using shared task files and worktree isolation.

## When to Use

- Setting up parallel agent execution
- Managing multiple concurrent workflows
- Scaling beyond single-agent patterns
- Building task queue systems

## Core Concept

Agents share a task file that acts as a coordination mechanism:

```markdown
## To Do
- [ ] Task A
- [ ] Task B

## In Progress
- [🟡 abc123] Task C - being worked on

## Done
- [✅ def456] Task D - completed
```

## Task File Format

`tasks.md`:

```markdown
# Tasks

## Git Worktree {worktree-name}

## To Do
[] Pending task description                           # Available
[⏰] Blocked task (waits for above)                   # Blocked
[] Task with #opus tag                                # Model override
[] Task with #adw_plan_implement tag                  # Workflow override

## In Progress
[🟡, adw_12345] Task being processed                  # Claimed by agent

## Done
[✅ abc123, adw_12345] Completed task                 # Commit hash saved
[❌, adw_12345] Failed task // Error reason           # Error captured
```

## Status Markers

| Marker | Meaning | State |
| --- | --- | --- |
| `[]` | Pending | Available for pickup |
| `[⏰]` | Blocked | Waiting for previous |
| `[🟡, {id}]` | In Progress | Being processed |
| `[✅ {hash}, {id}]` | Complete | Finished successfully |
| `[❌, {id}]` | Failed | Error occurred |

## Tag System

Tags modify agent behavior:

| Tag | Effect |
| --- | --- |
| `#opus` | Use Opus model |
| `#sonnet` | Use Sonnet model |
| `#adw_plan_implement` | Complex workflow |
| `#adw_build` | Simple build workflow |

## Implementation Architecture

```text
┌─────────────────────────────────────────┐
│            CRON TRIGGER                  │
│  (polls tasks.md every N seconds)        │
└─────────────────┬───────────────────────┘
                  │
        ┌─────────┼─────────┐
        │         │         │
        v         v         v