AI Skills
Headroom
A context-compression skill that reduces token usage by intelligently summarizing logs, documents, tool outputs, and large context before passing it to an LLM.
Official reviewHeadroom compresses context before it reaches the model — summarizing logs, documents, and tool output so you spend tokens on signal instead of bloat.
Pairs naturally with cost-conscious workflows; a practical lever when long tool outputs or documents keep blowing up your context window.
Sign in to join the discussion.