functional-area-resolver

To compress a large agent routing file, group individual skill entries into functional-area dispatchers. This hierarchical routing pattern reduces routing table size by up to 50% while maintaining high dispatching accuracy for LLM-based architectures.

A functional-area dispatcher is a two-layer routing pattern where the main file routes tasks to a functional area, which then handles specific sub-skills. It optimizes context budget by replacing long lists of individual skills with grouped dispatchers.

You need to reorganize your agent routing table when it exceeds context budget limits, such as growing beyond 200 rows and consuming 25KB of context. Reorganizing into 10-15 functional areas reclaims space for actual task execution.

Compress an AGENTS.md file by applying the dispatcher-aware prompt pattern to group entries into functional areas. Structural verification is performed via the provided harness to ensure the compressed routing file maintains proper dispatching logic.

The resolver targets LLM-based agent architectures where routing tables exceed context budget limits. While it can process smaller tables, the context optimization benefits are specifically designed for large files consuming excessive context space.

Compressing routing tables into functional areas requires adherence to the dispatcher-aware prompt pattern and structural verification via the provided harness. The two-layer routing approach may introduce slight routing indirection compared to flat routing tables.