Multi-Agent Task Decomposition Framework

A complete framework for decomposing complex tasks into subtasks that multiple agents can execute in parallel. Includes: task decomposition prompts that produce structured DAGs, dependency graph schemas, agent capability matching logic, result aggregation patterns, and conflict resolution strategies. Works with any multi-agent setup. Includes example decompositions for common scenarios: full-stack feature development, codebase migration, research reports, and security audits. The secret sauce is the decomposition prompt — it consistently produces clean, non-overlapping subtasks with explicit input/output contracts.