All notable changes to this project will be documented here.
Format based on Keep a Changelog. Versioning follows Semantic Versioning.
- Full 5-phase optimization pipeline: Discovery → Analysis → Report → Apply → Before/After
- 9 analysis detectors: duplication, misplacement, bloat, stale content, contradictions, token budget audit, best-practices comparison, missing repo index, security scan
- Three modes: interactive (summary-then-batch), auto (backup-then-apply), analyze (read-only)
- Restore mode: undo all changes from latest backup
- Backup with scope-matched paths (
~/.claude/backups/for global,.claude-md-backup/for project) - "What Claude Sees" token simulator per working directory
- Unicode bar chart token reports with before/after comparison
- Migration mode for splitting a monolithic CLAUDE.md into a proper hierarchy
- Auto-generate repo index for monorepos
- Live best-practices research via
--researchflag (fetches Anthropic docs) - Weekly automated sync of best-practices reference from Anthropic documentation
- Security scan detector (2i): flags accidentally committed secrets, API keys, internal URLs, credentials, and PII in CLAUDE.md files with line numbers and safe replacement suggestions. Security findings always appear first in the report.