ctx-audit

You audit CLAUDE.md token usage by running a step-based evaluation that checks token efficiency, hierarchy depth, and information redundancy across CLAUDE.md, CONTEXT.md, and AGENTS.md files to produce actionable recommendations.

Context architecture redundancy is duplicated information across CLAUDE.md, CONTEXT.md, and AGENTS.md layers. You reduce it by evaluating hierarchy depth and applying context distribution checks to eliminate bloat and improve token efficiency.

Yes, the audit applies to repositories with CLAUDE.md, and also scales to those utilizing CONTEXT.md and AGENTS.md, measuring context coverage and guiding distribution across multiple layers when available.

The best way to measure token efficiency is through a structured audit that evaluates context depth, performs redundancy checks across file layers, and outputs a prioritized improvement plan for lean context distribution.

Your CLAUDE.md context feels bloated due to information redundancy and suboptimal context distribution. An audit evaluates token usage and hierarchy depth to identify specific areas for reducing bloat and improving efficiency.

Auditing large repositories requires comprehensive context coverage measurement across distributed files, and limitations arise when the hierarchy depth and redundancy checks cannot fully map fragmented context layers into a unified operational roadmap.