Start Rapid

Rapidly initiate task orchestration workflows using the three-agent system to decompose requirements, analyze dependencies, and create a hyper-efficient execution plan.

When to Use This Command

Run this command when...

• You have a set of tasks or requirements and want to generate an optimized execution plan with dependency ordering
• You are beginning a new project phase and need the orchestration system to decompose work into actionable steps
• You want the three-agent system (orchestrator, decomposer, dependency-analyzer) to plan your workflow
• A requirements document or task list needs to be converted into a structured execution plan with parallel paths
• You need to bootstrap the task-orchestration directory structure for a new day or sprint

Quick Start

Copy
# .claude/commands/start-rapid.md
---
name: Start Rapid
description: Bootstrap orchestration with task decomposition and dependency analysis
command: true
---
Start orchestration for: $ARGUMENTS

Activate the three-agent system:
1. Task Orchestrator: Parse requirements and clarify scope
2. Task Decomposer: Break into atomic, actionable tasks
3. Dependency Analyzer: Map dependencies and find parallel paths

Copy
# Invoke the command
claude "/start-rapid build a user authentication system with OAuth and MFA"

# Expected output
# > Activating orchestration system...
# > Creating: task-orchestration/2026-03-15/
# > Task Decomposer: Breaking down requirements...
# >   TASK-001: Set up authentication database schema
# >   TASK-002: Implement JWT token service
# >   TASK-003: Build OAuth2 provider integration
# >   TASK-004: Create MFA enrollment flow
# >   TASK-005: Write authentication middleware
# >   TASK-006: Add login/logout API endpoints
# >   TASK-007: Create integration test suite
# > Dependency Analyzer: Mapping execution order...
# >   Parallel group 1: TASK-001, TASK-002
# >   Sequential: TASK-003 (after TASK-001)
# >   Parallel group 2: TASK-004, TASK-005 (after TASK-002)
# >   Sequential: TASK-006 (after TASK-003, TASK-005)
# >   Final: TASK-007 (after all)
# > Execution plan saved. Ready to begin.

Core Concepts

Concept	Description
Three-Agent System	Orchestrator, Decomposer, and Dependency Analyzer work in sequence to create plans
Task Decomposition	Breaks high-level requirements into atomic, independently executable tasks
Dependency Graph	Maps which tasks block others and identifies opportunities for parallel execution
Directory Scaffolding	Creates date-stamped orchestration directories with task files and metadata
Parallel Path Detection	Finds tasks with no mutual dependencies that can be executed simultaneously

Start Rapid Three-Agent Flow
==============================

  Requirements
       |
       v
  [Orchestrator]     Clarify scope, confirm understanding
       |
       v
  [Decomposer]      Break into TASK-001, TASK-002, ...
       |
       v
  [Dependency        Map: TASK-002 depends on TASK-001
   Analyzer]         Find: TASK-003, TASK-004 can run parallel
       |
       v
  Execution Plan     Ordered groups with parallel paths
       |
       v
  task-orchestration/2026-03-15/
    tasks.md | dependencies.md | plan.md

Configuration

Parameter	Description	Default	Example	Required
$ARGUMENTS	Requirements description or path to requirements file	none	"build auth system"	Yes
max_task_depth	Maximum decomposition depth for sub-tasks	2	3	No
parallel_limit	Maximum tasks allowed in a parallel execution group	4	2	No
output_dir	Custom directory for orchestration files	task-orchestration/{date}	"sprint-14/"	No
auto_start	Immediately begin executing after planning	false	true	No

Best Practices

1. Provide detailed requirements upfront -- The more context you give, the better the decomposition. Include constraints, priorities, and acceptance criteria in your arguments.

2. Review the dependency graph before executing -- The analyzer may miss implicit dependencies or create overly sequential plans. Verify the parallel groups make sense for your project.

3. Set appropriate parallel limits -- Running four tasks in parallel works well for independent modules but may cause conflicts when tasks modify shared files. Adjust based on coupling.

4. Use file paths for complex requirements -- For large requirement documents, pass a file path instead of inline text: "/start-rapid @docs/requirements.md"

5. Start each day with a fresh orchestration -- Use date-stamped directories to keep orchestration state organized and avoid mixing work from different sessions.

Common Issues

Over-decomposition creates too many tasks: Very detailed requirements can produce dozens of micro-tasks. Increase max_task_depth: 1 to keep tasks at a higher granularity, or merge related tasks manually.

Dependency analyzer misses implicit dependencies: Tasks that share database tables or API contracts may not have explicit code dependencies but still conflict. Add manual dependency annotations after the initial analysis.

Directory structure already exists: If you run start-rapid twice on the same day, it may conflict with existing orchestration files. The command appends a sequence number or prompts to overwrite.