intelligent-do

Automate end-to-end task execution by using a single command that handles context-aware planning, specification generation, and wave-based orchestration. The Skill detects your project state, assesses complexity, and coordinates execution via Serena MCP as a persistent memory backend, eliminating manual handoffs between planning and delivery stages.

Yes. The Skill works with both greenfield projects and existing codebases by automatically loading project context and checkpoints. It adapts its planning and execution strategy based on whether you're starting fresh or returning to a cached workflow with prior context.

Wave-based execution breaks complex tasks into coordinated batches that execute in sequence, allowing the system to validate progress and adjust context between waves. This approach handles tasks from simple to complex, including first-time explorations and research-driven workflows, while maintaining coherent command orchestration.

Serena MCP serves as a persistent memory backend that stores project state, previous execution results, and learned complexity patterns. This integration allows the Skill to retrieve cached context on return visits, eliminating redundant analysis and enabling faster, context-aware re-execution of related workflows.

No. The Skill includes optional spec-analysis that automatically generates specifications based on your project state and task complexity. Manual specification is available when you need fine-grained control, but the system can derive and refine specs without explicit input.

The Skill performs complexity assessment upfront and adapts its planning strategy accordingly. For complex tasks, it breaks work into smaller, manageable waves with intermediate checkpoints, ensuring even high-complexity tasks complete reliably with proper context coordination across stages.