rlm-long-context
Installation
SKILL.md
RLM Long-Context Processing (Experimental)
For production use, prefer the
rlmskill which uses the fleet-rlm package with Daytona sandboxes.Use this skill for evaluating alternative RLM strategies, researching optimization techniques, or comparing approaches.
Scripts & References
| Resource | Purpose |
|---|---|
scripts/orchestrate.py |
Main orchestrator with all optimizations |
scripts/rank_chunks.py |
Query-guided chunk selection (5-10x speedup) |
scripts/semantic_chunk.py |
Content-aware chunking by boundaries |
scripts/cache_manager.py |
Result caching for repeated queries |
scripts/codebase_concat.py |
Concatenate codebase files for processing |
| references/advanced-techniques.md | Query-guided selection, semantic chunking, adaptive sizing, map-reduce, streaming, caching |
| references/codebase-processing.md | Whole-codebase analysis: concatenation, code chunking, file selection, query types |