Concurrent Coding Task Allocation and Distribution Interface

[justin4957]

Description

Create an interactive interface for allocating and distributing coding tasks to multiple AI providers concurrently. This allows users to break down complex coding projects and assign different subtasks to different providers, monitoring their progress in real-time.

Goals

• Enable users to define and allocate coding tasks to specific providers
• Support task decomposition (breaking large tasks into subtasks)
• Allow dynamic task assignment based on provider capabilities
• Track task dependencies and execution order
• Monitor task completion status across all providers

Tasks

• Implement task allocation interface:
  • Command to define a new coding task
  • Assign tasks to specific providers or use auto-allocation
  • Support task templates (e.g., "implement feature X", "write tests for Y")
  • Allow task prioritization
• Create task decomposition system:
  • Parse complex tasks into subtasks
  • Suggest optimal provider for each subtask
  • Handle task dependencies (task B depends on task A)
• Build task queue management:
  • Queue visualization showing pending/active/completed tasks
  • Support task reordering and cancellation
  • Retry failed tasks with different providers
• Implement provider capability matching:
  • Define provider strengths (e.g., OpenAI for algorithms, Claude for refactoring)
  • Auto-assign tasks based on capabilities
  • Load balancing across providers
• Add task progress tracking:
  • Show which provider is working on which task
  • Display estimated completion time
  • Track tokens used per task per provider
• Support task result aggregation and comparison

Interactive Commands

> allocate "Implement user authentication system" to openai,anthropic
> split "Create REST API" into 3 subtasks
> assign task-1 to anthropic
> assign task-2,task-3 to openai
> queue status
> retry task-2 with local
> merge results from task-1,task-2

Files

• lib/multi_agent_coder/task/allocator.ex (new) - Task allocation logic
• lib/multi_agent_coder/task/decomposer.ex (new) - Break down complex tasks
• lib/multi_agent_coder/task/queue.ex (new) - Task queue management
• lib/multi_agent_coder/task/tracker.ex (new) - Track task progress
• lib/multi_agent_coder/cli/command.ex - Add allocation commands
• test/multi_agent_coder/task/*_test.exs (new)

Related Issues

• Works with Concurrent Provider Display with Real-time Split View #11 for displaying concurrent task execution
• Depends on provider implementations
• Integrates with Implement Session Persistence #5 for task persistence

Example Workflow

> allocate "Build a Phoenix API with authentication" to all
Task decomposed into:
  1. Create Phoenix project structure [auto → anthropic]
  2. Implement user schema and migrations [auto → openai]  
  3. Add authentication logic [auto → anthropic]
  4. Create API endpoints [auto → openai]
  5. Write tests [auto → all]

> start
[Shows concurrent execution across providers...]