/parallel-tasks

allowed-tools: Task, TodoWrite, mcp__spec-workflow__spec-list, mcp__spec-workflow__manage-tasks, mcp__spec-workflow__spec-status, mcp__linear__list_issues, mcp__linear__list_projects, Bash, Read, LS, Grep, Glob

Parallel Tasks Execution

Deploy specialized AI agents to work on multiple tasks simultaneously while maintaining coordination and task tracking. This command can reduce development time by 60-80% through intelligent agent coordination and task parallelization.

Instructions

1. **Prerequisites and Setup**

• Verify project has active task management system (spec-workflow, GitHub, Linear)
• Ensure no conflicting processes are running (other builds, deploys)
• Check system resources available for parallel agent execution
• Review available command arguments:
• `--system <TYPE>` - Task management system: spec-workflow, github, linear, manual
• `--project-path <PATH>` - Absolute path to project root (default: current directory)
• `--spec-name <NAME>` - Specification name (required for spec-workflow system)
• `--auto-assign` - Automatically assign tasks to optimal agents (recommended)
• `--max-agents <NUMBER>` - Maximum parallel agents (default: 4, range: 1-8)
• `--dry-run` - Preview task assignments without execution
• `--conflict-resolution <MODE>` - Handle conflicts: abort, merge, manual (default: abort)
• `--resource-limit <LEVEL>` - Resource usage: low, medium, high (default: medium)
• `--help` - Display comprehensive help documentation

2. **Preflight Validation**

• **Help Flag Detection**: If `--help` flag is present, display comprehensive help documentation instead of executing
• **System Connectivity Check**:
• **Spec-Workflow**: Verify MCP server connection: `mcp__spec-workflow__spec-list`
• **GitHub**: Check gh CLI access: `gh auth status`
• **Linear**: Test Linear MCP connection: `mcp__linear__list_projects`
• **Project Path Validation**: Verify project directory exists, check for git repository
• **Resource Assessment**: Check available system resources and validate max-agents parameter (1-8)

3. **Execute Parallel Task Processing**

**Phase 1: Task Discovery and Analysis**

**System Integration:**

• Connect to specified task management system
• For spec-workflow: Use `mcp__spec-workflow__spec-list` to find specifications
• For spec-workflow with spec-name: Use `mcp__spec-workflow__manage-tasks` with action="list"
• Extract all pending/available tasks with their details

**Task Analysis:**

# Example for spec-workflow
1. Get task list with dependencies and file information
2. Analyze task descriptions for type classification
3. Identify file modification conflicts between tasks
4. Group tasks by specialization area (testing, deployment, development)

**Parallelization Assessment:**

• Tasks that modify same files → Sequential execution required
• Tasks with explicit dependencies → Respect dependency order
• Independent tasks of different types → Prime candidates for parallelization
• Large tasks → Consider if they can be subdivided

**Phase 2: Agent Assignment Strategy**

**Task Categorization:**

Create task groups based on content analysis:

• **Testing Tasks**: Keywords: test, spec, coverage, validation, qa
• **Performance Tasks**: Keywords: benchmark, optimization, performance, metrics
• **Deployment Tasks**: Keywords: deploy, ci/cd, script, infrastructure
• **Development Tasks**: Keywords: implement, feature, refactor, bug fix
• **Documentation Tasks**: Keywords: docs, readme, api, documentation

**Agent Matching Matrix:**

Testing Tasks:
  - test-automator: Unit tests, test configuration, coverage
  - qa-expert: Quality assurance, validation, acceptance testing
  - playwright-test-architect: E2E tests, integration testing

Performance Tasks:
  - performance-engineer: Benchmarking, optimization, metrics

Deployment Tasks:
  - deployment-engineer: CI/CD, scripts, infrastructure automation

Development Tasks:
  - typescript-pro: TypeScript development, refactoring
  - react-pro: React components, frontend development
  - python-pro: Python backend, API development
  - golang-pro: Go services, system programming

Documentation Tasks:
  - documentation-expert: Technical docs, guides, architecture
  - api-documenter: API documentation, examples, SDKs

**Conflict Detection:**

• File-level conflict analysis (no two agents modify same files)
• Resource conflict prevention (database, external services)
• Dependency ordering (Task A must complete before Task B)

**Phase 3: Parallel Agent Deployment**

**Pre-Deployment Validation:**

• Confirm no file conflicts in task assignments
• Verify all required tools/dependencies are available
• Set up coordination mechanism for status tracking
• Create TodoWrite list to track overall progress

**Agent Deployment Process:**

For each selected agent and task group:

1. **Status Update**: Set tasks to "in-progress" in source system

# Example for spec-workflow
mcp__spec-workflow__manage-tasks --action set-status --task-id X --status in-progress

2. **Agent Prompt Template**:

Task {agent_type} agent to handle: {task_descriptions}

**CRITICAL WORKFLOW INTEGRATION**:
1. BEFORE starting: Task status already set to in-progress
2. Complete assigned work with full implementation
3. AFTER completion: Set task status to completed using source system

**Assigned Tasks**: {task_list}
**Files to Modify**: {file_list}
**Requirements**: {requirements}
**Integration Points**: {dependencies}
**Coordination**: Report progress and conflicts immediately

3. **Parallel Launch**: Use Task tool to deploy multiple agents simultaneously

# Deploy up to max-agents simultaneously
Task(subagent_type=agent_type, description=task_summary, prompt=full_prompt)

**Phase 4: Coordination and Monitoring**

**Real-time Coordination:**

• Track agent progress through TodoWrite updates
• Monitor for file co