context-audit

To optimize context density and instruction efficiency in CLAUDE environments, conduct an audit that measures context surfaces, detects redundant patterns like CLAUDE.md echo, and categorizes content into priority tiers to improve attention allocation.

Attention dilution in AI system instructions is caused by low instruction density and redundant patterns such as cross-component cloning and agent-skill overlap, which this audit identifies to help you target compression and improve efficiency.

You detect redundant patterns across CLAUDE.md and hooks by running a context audit that scans for hook echo, cross-layer duplication, and agent-skill overlap, categorizing the findings into priority tiers for targeted compression.

No, the context audit requires no external dependencies to analyze CLAUDE context surfaces, measure line counts, and generate efficiency recommendations using its built-in scripts and references.

You should audit context surfaces for path-scoping and attention budget utilization when instruction density drops or redundancy increases, allowing the audit to categorize sections by importance and suggest compression targets.

The best way to categorize AI context content into priority tiers for attention allocation is to run a full context audit, which measures effective instructions against total lines and identifies path-scoping opportunities to maximize attention budget utilization.