context-budget

Auditing token consumption in Claude Code involves scanning agents, skills, MCP servers, rules, and CLAUDE.md files to estimate and categorize token usage. This process generates a budget report flagging bloated or redundant components with actionable optimization recommendations to reclaim context space.

Claude Code sessions slow down when bloated agents, redundant skills, oversized MCP servers, and verbose CLAUDE.md files consume excessive context window tokens. Running a token budget audit identifies heavy, low-value items and provides prioritized recommendations to trim token waste and restore performance.

Yes, you can check available context capacity by running a token budget audit before expanding your Claude Code system. The audit performs file scanning and token estimation across all loaded components, generating a report that verifies whether you have enough remaining context capacity for new additions.

To identify wasting context tokens, the audit scans loaded agents, skills, rules, and MCP configurations, estimating token consumption and categorizing components by necessity. It flags heavy, redundant, or oversized items, highlighting specific bloated MCP servers and agents for optimization or removal.

The best way to reduce token waste from verbose CLAUDE.md files is to run a context budget audit that scans and estimates file token consumption. The resulting budget report categorizes these files by necessity and provides prioritized, actionable optimization recommendations to trim verbose content.

Running a context budget audit requires no specific dependencies or components. The Skill independently performs file scanning, token estimation, and categorization across agents, skills, MCP servers, and rules to generate its budget report without needing external tools.