context-budget

To audit context window usage in Claude Code, run a context-budget audit to inventory agents, skills, MCP servers, and rules. This estimates token consumption across components and flags bloated files consuming excessive context headroom.

Claude Code sessions slow down and consume excessive tokens when multiple components accumulate bloat in the context window. A structured context audit identifies redundant rules and heavy files to reclaim context headroom and improve performance.

The best way to identify token bloat across MCP servers and rules is to perform a structured context audit. This classifies components into always-needed, sometimes-needed, and rarely-needed categories to guide on-demand activation and produce token-savings recommendations.

Yes, you can generate a token-savings report for a large Mullai deployment by running a context-budget audit. This surfaces top token-savings opportunities, detects redundant components, and provides guidance to de-scope unused skills to reduce context usage.

You should run a context audit to reclaim token space when Claude Code sessions accumulate excessive context from multiple components, leading to slower performance and higher costs. The audit detects issues and outputs a prioritized list of optimizations.