brewdocmemorylisted

# Memory Optimizer Optimizes Claude Code memory files through 4 interactive steps. > **No `context: fork`** — must run in main conversation to spawn agents. ## Phase 0: Load Context 0. **Determine memory directory (`$MEMORY_DIR`):** ```bash CUSTOM_DIR=$(cat .claude/settings.json 2>/dev/null | jq -r '.autoMemoryDirectory // empty') if [ -n "$CUSTOM_DIR" ]; then MEMORY_DIR="$(git rev-parse --show-toplevel)/$CUSTOM_DIR" else MEMORY_DIR=~/.claude/projects/<hash>/memory fi ``` Read `.claude/settings.json` (if exists) → extract `autoMemoryDirectory`. If set → resolve as `<git-root>/<autoMemoryDirectory>`. If not set → use legacy `~/.claude/projects/<hash>/memory/` glob pattern. 1. Glob all memory files: `$MEMORY_DIR/*.md` (or `~/.claude/projects/**/memory/*.md` for legacy) 2. Read `~/.claude/CLAUDE.md` and project `CLAUDE.md` (if exists) 3. Glob `.claude/rules/*.md` — read all project rules 4. Read `~/.claude/rules/*.md` — read all global rules Build context map: ``` memory_dir: $MEMORY_DIR memory_files: [paths] claude_md_sections: [sections] rules_files: [paths with content] ``` ## Step 1: Analysis — Remove Duplicates (Interactive) **Goal:** Find memory entries that duplicate content already in CLAUDE.md or rules. 1. Spawn `Explore` agent to cross-reference all loaded files 2. Identify entries where: - Same rule already in CLAUDE.md - Same pattern already in a rules file - Contradicts CLAUDE.md (CLAUDE.md wins) 3. Show analysi