claude-md-improver

To audit CLAUDE.md files, scan project roots, packages, and modules to evaluate context, governance, and instruction currency against a standard rubric. This generates a delta report with actionable recommendations and diffs for targeted updates.

CLAUDE.md documentation drifts when project architecture, commands, and governance instructions evolve without corresponding markdown updates. This degrades AI session quality by providing outdated or incomplete project context to the model.

Standardize frontmatter and architecture clarity by enforcing strict frontmatter expectations during a repository scan. The audit analyzes commands and architecture details, then proposes targeted markdown diffs after your approval.

Yes, you can generate a quality report for per-package and per-module markdown documentation. The audit discovers all CLAUDE.md variants across the repository and assesses them against defined criteria including currency and governance.

The best way to maintain consistent project context is to regularly audit CLAUDE.md files against a standardized rubric. This identifies quality gaps and produces targeted diffs to keep documentation aligned with the codebase.

A limitation of automating markdown documentation quality reviews is that proposed updates require user approval before application. The audit enforces strict frontmatter expectations but focuses on context, governance, and currency rather than arbitrary formatting.