recursive-decomposition

Recursive decomposition handles tasks that exceed context limits by splitting large problems into smaller subproblems and orchestrating sub-agents to combine partial results into a synthesized final answer.

Recursive decomposition manages multi-file research by applying recursive self-invocation to split context-heavy inputs into staged explorations, allowing sub-agents to process chunks and synthesize aggregated results.

Yes, you can analyze cross-cutting concerns like security across 10+ files by applying recursive task decomposition to split the codebase analysis into manageable subproblems and aggregating the partial findings.

The best way to orchestrate sub-agents for context-aware aggregation is to use recursive self-invocation to delegate structured subproblems, gather partial analyses, and synthesize a final answer through robust staged exploration.

Recursive self-invocation for task planning is limited by its reliance on structured delegation and controlled context management, making it unsuitable for simple tasks that fit within single-prompt limits or require no multi-file synthesis.

No specific dependencies are required to automate task decomposition; the skill operates independently by using recursive self-invocation to manage context and coordinate sub-agents for scalable project planning.