parallel-task-format

Compact YAML format for defining parallel task specifications with scope, boundaries, and agent assignments. Use when creating task files for parallel development.

16 stars

Best use case

parallel-task-format is best used when you need a repeatable AI agent workflow instead of a one-off prompt.

Compact YAML format for defining parallel task specifications with scope, boundaries, and agent assignments. Use when creating task files for parallel development.

Teams using parallel-task-format should expect a more consistent output, faster repeated execution, less prompt rewriting.

When to use this skill

  • You want a reusable workflow that can be run more than once with consistent structure.

When not to use this skill

  • You only need a quick one-off answer and do not need a reusable workflow.
  • You cannot install or maintain the underlying files, dependencies, or repository context.

Installation

Claude Code / Cursor / Codex

$curl -o ~/.claude/skills/parallel-task-format/SKILL.md --create-dirs "https://raw.githubusercontent.com/diegosouzapw/awesome-omni-skill/main/skills/development/parallel-task-format/SKILL.md"

Manual Installation

  1. Download SKILL.md from GitHub
  2. Place it in .claude/skills/parallel-task-format/SKILL.md inside your project
  3. Restart your AI agent — it will auto-discover the skill

How parallel-task-format Compares

Feature / Agentparallel-task-formatStandard Approach
Platform SupportNot specifiedLimited / Varies
Context Awareness High Baseline
Installation ComplexityUnknownN/A

Frequently Asked Questions

What does this skill do?

Compact YAML format for defining parallel task specifications with scope, boundaries, and agent assignments. Use when creating task files for parallel development.

Where can I find the source code?

You can find the source code on GitHub using the link provided at the top of the page.

SKILL.md Source

# Task Specification Format

Compact YAML format for parallel task files in `parallel/TS-XXXX-slug/tasks/`.

## Complete Task Example

```yaml
---
id: task-001
component: users
wave: 1
deps: []
blocks: [task-004, task-005]
agent: python-experts:django-expert
skills: [python-experts:python-style, python-experts:django-dev, python-experts:django-api]
tech_spec: TS-0042
contracts: [contracts/types.py, contracts/api-schema.yaml]
---
# task-001: User Management

## Scope
CREATE: apps/users/{models,views,serializers,urls}.py, apps/users/tests/*.py
MODIFY: config/urls.py
BOUNDARY: apps/orders/*, apps/products/*, apps/*/migrations/*

## Requirements
- User model with email authentication
- UserSerializer with explicit fields
- UserViewSet (list, retrieve, create, update)
- Email uniqueness validation

## Checklist
- [ ] Model matches UserDTO in contracts/types.py
- [ ] API matches /api/users/* in contracts/api-schema.yaml
- [ ] pytest apps/users/ passes
- [ ] mypy apps/users/ passes
- [ ] ruff check apps/users/ passes
- [ ] No files modified outside scope
```

## YAML Frontmatter Fields

| Field | Required | Description |
|-------|----------|-------------|
| `id` | Yes | Task identifier (task-NNN or task-NNN-component) |
| `component` | Yes | System component name |
| `wave` | Yes | Dependency wave number (1, 2, 3...) |
| `deps` | Yes | Task IDs this depends on (empty list `[]` if none) |
| `blocks` | No | Task IDs this blocks (optional) |
| `agent` | Yes | Recommended agent type |
| `skills` | Yes | Skills the agent should invoke (list from agent-skills-mapping.yaml) |
| `tech_spec` | No | Tech Spec ID (if applicable) |
| `contracts` | Yes | Contract files to reference (relative paths) |

**Note:** `skills` are stored in task files so prompt generation can include them in `=== REQUIRED SKILLS ===` sections. They are NOT stored in manifest.json.

## Scope Section Format

Use compact notation with three directives:

### CREATE
Files to create (use glob patterns). **A file can only be in CREATE for ONE task.**
```
CREATE: apps/users/{models,views,serializers,urls}.py, apps/users/tests/*.py
```

### MODIFY
Existing files to modify. Use **scoped syntax** for parallel modifications:

**Unscoped** (whole file - only ONE task per wave can use this):
```
MODIFY: config/urls.py, config/settings.py
```

**Scoped** (specific section - multiple tasks in same wave can modify different scopes):
```
MODIFY: apps/users/models.py::User.save          # Owns User.save method
MODIFY: apps/users/models.py::User.clean         # Different task owns User.clean
MODIFY: apps/users/views.py::UserViewSet         # Owns entire class
MODIFY: config/urls.py::urlpatterns              # Owns urlpatterns list
```

**Scoped syntax rules:**
- `file.py::ClassName` - owns entire class
- `file.py::function_name` - owns entire function
- `file.py::ClassName.method` - owns specific method
- Scopes must NOT overlap (no nesting like `::Class` and `::Class.method` in same wave)

### BOUNDARY
Files NOT to touch (owned by other tasks):
```
BOUNDARY: apps/orders/*, apps/products/*, apps/*/migrations/*
```

## Task Naming Convention

```
task-{number}-{component}.md

Examples:
- task-001-users.md
- task-002-products.md
- task-003-orders.md
- task-004-api.md
- task-005-integration.md
```

## Agent Type Selection

| Task Files | Agent | Description |
|------------|-------|-------------|
| `apps/*/models.py`, `apps/*/views.py` | `python-experts:django-expert` | Django models, views, serializers |
| `api/*.py`, `routers/*.py` | `python-experts:fastapi-expert` | FastAPI endpoints |
| `src/components/*.tsx` | `frontend-experts:react-typescript-expert` | React components |
| `**/test_*.py`, `**/tests/*.py` | `python-experts:python-testing-expert` | Python tests |
| `*.spec.ts`, `*.test.tsx` | `frontend-experts:playwright-testing-expert` | TypeScript/E2E tests |
| `terraform/`, `docker-compose.yml` | `devops-data:devops-expert` | Infrastructure |
| Integration, architecture | `devops-data:cto-architect` | Cross-cutting concerns |

## Contract References

Contracts are in the same parallel directory:
```
parallel/TS-0042-slug/
  contracts/
    types.py        # Reference as: contracts/types.py
    api-schema.yaml # Reference as: contracts/api-schema.yaml
  tasks/
    task-001-users.md
```

## Wave Dependencies

Tasks in Wave N can only depend on tasks in Waves 1 to N-1:

```
Wave 1: task-001, task-002 (no dependencies, run in parallel)
Wave 2: task-003 (depends on task-001, task-002)
Wave 3: task-004 (depends on task-003)
```

### deps vs blocks

- `deps`: Tasks that MUST complete before this task starts
- `blocks`: Tasks that CANNOT start until this task completes

Both express the same relationship from different perspectives:
```yaml
# task-001
blocks: [task-003]

# task-003
deps: [task-001]
```

## Requirements Section

Clear, actionable requirements:

```markdown
## Requirements
- Implement `User` model with fields: `email`, `username`, `password`, `is_active`
- Create `UserSerializer` with all User fields (hide password)
- Implement `UserViewSet` with: list, retrieve, create, update
- Add email validation and uniqueness constraint
- Test coverage: minimum 85%
```

## Checklist Section

Verification criteria:

```markdown
## Checklist
- [ ] Model matches DTO in contracts/types.py
- [ ] API matches schema in contracts/api-schema.yaml
- [ ] pytest apps/users/ passes
- [ ] mypy apps/users/ --strict passes
- [ ] Coverage >= 85%
- [ ] No files modified outside scope
```

## Why Compact Format?

1. **Token efficiency**: Less tokens for agent context
2. **Faster parsing**: YAML frontmatter is standard
3. **Clear boundaries**: Scope section is scannable
4. **Actionable checklist**: Verification is explicit

## Validation Rules

Before using tasks:

- [ ] Every task has unique `id`
- [ ] Every task has `agent` assigned
- [ ] Every task has `skills` list (from agent-skills-mapping.yaml)
- [ ] Every task has `contracts` referenced
- [ ] Every task has BOUNDARY section
- [ ] No circular dependencies in `deps`
- [ ] Wave numbers are sequential (1, 2, 3...)
- [ ] Wave 1 tasks have `deps: []`

## Output Format

The Output Format JSON block is **not included in task files or generated prompts**. It's managed via system prompt by the external execution tool (`cpo` orchestrator).

This ensures consistent JSON output format across all agents without duplicating the schema in every prompt.

Related Skills

parallel-workflows

16
from diegosouzapw/awesome-omni-skill

Optimizes parallel execution of multiple tasks. Use when user mentions 並列で実行, 同時にやって, まとめてやって, run in parallel, do these together. Do NOT load for: 単一タスク, 順次実行が必要な作業, 依存関係のあるタスク.

Meta Dispatcher & Task Orchestrator

16
from diegosouzapw/awesome-omni-skill

PRD 驱动的任务调度与技能管理专家。接收完整 PRD/需求文档,负责拆解业务、选择技术栈、路由到合适的专业 Skill,并维护从方案到落地的全流程。

manage-tasks

16
from diegosouzapw/awesome-omni-skill

Manage implementation tasks with sequential sub-steps within a plan

history-and-next-task-rules

16
from diegosouzapw/awesome-omni-skill

Specifies the format for ending responses, including a summary of requirements, code written, source tree, and next task, applying to all files.

format-figure

16
from diegosouzapw/awesome-omni-skill

Reformats a figure component file to conform to the standard section order and naming conventions defined in the figure guidelines.

format-euler-code

16
from diegosouzapw/awesome-omni-skill

Enforce the Euler repository C++ formatting and style rules from AGENTS.md. Use when writing, editing, or reviewing Euler solution code so it matches required includes, types, layout, namespaces, and I/O conventions.

Fixed Video Format (9:16)

16
from diegosouzapw/awesome-omni-skill

Fixed 1080x1920 pixel video format with percentage-based positioning. Use this when laying out video compositions, positioning elements on the canvas, or calculating dimensions. All videos render at exactly 9:16 aspect ratio for TikTok/Instagram Reels.

code_formatter

16
from diegosouzapw/awesome-omni-skill

Otomatik kod formatlama, Prettier/ESLint entegrasyonu ve kod stil tutarlılığı rehberi.

claude-code-task

16
from diegosouzapw/awesome-omni-skill

Run Claude Code tasks in background with automatic result delivery. Use for coding tasks, research in codebase, file generation, complex automations. Zero OpenClaw tokens while Claude Code works.

architecture-format-extended

16
from diegosouzapw/awesome-omni-skill

Extended architecture templates with full examples. Imports architecture-format-core for base structure.

adhd-task-management

16
from diegosouzapw/awesome-omni-skill

ADHD-optimized task tracking and intervention system. Use when tracking tasks, detecting context switches, providing accountability interventions, or managing ADHD-specific productivity patterns for Ariel Shapira.

actionable-review-format-standards

16
from diegosouzapw/awesome-omni-skill

Standardized output format for code reviews with severity labels, file:line references, and fix code snippets. Use when generating review reports that need consistent, actionable feedback structure.