implement-orchestrator

Automating code generation for large designs is handled by grounding project docs and spawning Sonnet agents to parallelize implementation across subsystems. The orchestrator decomposes designs into agent-sized work units, tracks dependencies, and ensures production-ready code aligned with project conventions.

Agent orchestration for design-to-code implementation coordinates multiple Sonnet agents to translate design documents into production-ready code. It works by grounding the design, planning tasks with clear dependencies, and executing parallel work units to ensure alignment with project conventions.

Orchestrating multiple agents for code generation is needed for large designs with 20+ files or multiple subsystems. It requires careful dependency management and progress tracking to parallelize implementation effectively and ensure verified builds.

Orchestration aligns generated code with existing project conventions by grounding the implementation in design docs, CLAUDE.md, and the codebase. Sonnet agents use these conventions to produce production-ready code that passes verified builds and tests.

To start orchestrating agent implementation, provide the design document and start the orchestration. The process grounds the design and project docs, spawns Sonnet agents to work on independent subsystems, tracks progress, and supports cross-session resumes.

Agent orchestration supports resuming implementation across multiple sessions by tracking progress for each spawned Sonnet agent. This allows large designs with 20+ files to be implemented and verified over time without losing dependency management state.