rlm-skill

Recursive Language Model (RLM) pattern for processing large contexts. Use when dealing with massive files, logs, repos, or data that exceeds context windows. Trigger phrases: "analyze large file", "process huge log", "scan entire repo", "recursive context", "RLM", "context compilation", "unbounded context", "too big for context".

Show RLM token savings dashboard — how much context was saved by using RLM patterns instead of dumping raw data into the context window. Trigger: /rlm:stats