claude-slim 2.2.0

package/package.json

@@ -0,0 +1,50 @@
+{
+  "name": "claude-slim",
+  "version": "2.2.0",
+  "description": "Analyze and reduce Claude Code token overhead",
+  "type": "module",
+  "bin": {
+    "claude-slim": "./dist/cli.js"
+  },
+  "files": [
+    "dist",
+    "skills",
+    "README.md",
+    "LICENSE"
+  ],
+  "scripts": {
+    "build": "tsc",
+    "dev": "tsc --watch",
+    "test": "vitest run",
+    "test:watch": "vitest",
+    "prepublishOnly": "npm run build"
+  },
+  "keywords": [
+    "claude",
+    "claude-code",
+    "token",
+    "optimization",
+    "cleanup",
+    "skills",
+    "plugin"
+  ],
+  "author": "iops-leo",
+  "license": "MIT",
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/iops-leo/claude-slim.git"
+  },
+  "engines": {
+    "node": ">=18"
+  },
+  "dependencies": {
+    "commander": "^13.0.0",
+    "js-tiktoken": "^1.0.18"
+  },
+  "devDependencies": {
+    "@types/node": "^22.0.0",
+    "tmp-promise": "^3.0.3",
+    "typescript": "^5.7.0",
+    "vitest": "^4.1.4"
+  }
+}

package/skills/claude-slim/SKILL.md

@@ -0,0 +1,145 @@
+---
+name: claude-slim
+description: "Analyze and reduce Claude Code token overhead. Scans skills, plugins, memory files, and CLAUDE.md for bloat — then cleans up with user approval. Use when: /claude-slim, token optimization, reduce tokens, slim down, cleanup skills, token diet, save tokens, context diet"
+---
+
+# claude-slim — Token Overhead Reducer
+
+Analyze the user's Claude Code environment for token waste and perform non-destructive cleanup.
+
+## Subcommands
+
+- `/claude-slim` or `/claude-slim run` → full pipeline (scan → propose → execute → report)
+- `/claude-slim scan` → report only, no changes
+- `/claude-slim scan --json` → raw JSON output
+- `/claude-slim restore` → restore previously disabled items
+
+---
+
+## Phase 1 — Scan
+
+Run the CLI to collect environment data:
+
+```bash
+cd "${CLAUDE_PLUGIN_ROOT}" && node dist/cli.js scan --json
+```
+
+If `CLAUDE_PLUGIN_ROOT` is not set:
+```bash
+PLUGIN_DIR=$(find ~/.claude/plugins -path "*/claude-slim/dist/cli.js" -type f 2>/dev/null | head -1 | xargs dirname | xargs dirname)
+cd "$PLUGIN_DIR" && node dist/cli.js scan --json
+```
+
+If `node` is not available, fall back to the legacy bash scanner:
+```bash
+bash "${CLAUDE_PLUGIN_ROOT}/skills/claude-slim/scripts/scan.sh"
+```
+
+---
+
+## Phase 2 — Interpret & Present
+
+After getting the scan JSON, YOU must interpret and present results to the user. Do NOT just dump raw CLI output. Present a full diagnostic report in the user's language.
+
+### 2-1. Environment Snapshot Table
+
+Show a summary table:
+
+| 항목 | 수치 | 토큰 |
+|------|------|------|
+| 로컬 스킬 | N개 (XKB) | X tok |
+| 플러그인 | N개 (M 스킬) | ~X tok |
+| CLAUDE.md | XKB | X tok |
+| 메모리 파일 | N개 (XKB) | ~X tok |
+| **세션 시작 오버헤드** | | **~X tok** |
+
+### 2-2. Plugin Detail Table
+
+List each plugin with skill count and a judgment:
+
+| 플러그인 | 스킬 수 | 비고 |
+|----------|:-------:|------|
+| omc | 36 | 코어 플러그인. 유지 |
+| temp_local_... | 1 | **실패한 설치 잔여물. 삭제 대상** |
+
+Annotate each with status: actively used, possibly unused, or cleanup target. Flag `temp_local_*` entries as failed install remnants.
+
+### 2-3. Issue Analysis by Tier
+
+Group issues by tier and explain EACH one with context and recommendation:
+
+**Tier 1 — 즉시 정리 (위험 없음):**
+These are safe to remove with zero risk: broken symlinks, empty templates, .skill/ duplicates, temp_local_* cache. Pre-selected. Explain why each is safe.
+
+**Tier 2 — 정리 추천:**
+These are recommended but need user judgment. For each issue, explain:
+- What is it and why it's flagged
+- What happens if you remove it (safe? any side effects?)
+- How many tokens it saves
+
+Example: "frontend-design이 로컬과 플러그인에 둘 다 있습니다. 로컬 제거해도 플러그인 버전이 남으니 안전하게 제거 가능. ~823 tok 절감."
+
+**Tier 3 — 선택 사항 (사용자 판단):**
+These are large skills that cost tokens but might be in active use. For each:
+- Show size and token cost
+- Judge whether the user likely uses it (based on what it does)
+- Recommend: keep if active, disable if rarely used
+
+### 2-4. Recommended Actions
+
+End with a numbered action list, ordered by impact:
+1. What to do first (highest token savings, lowest risk)
+2. What to consider
+3. What to leave alone and why
+
+Show estimated total token savings if all recommended actions are taken.
+
+If subcommand is `scan`, stop here. Ask "정리할까요?" only for the full pipeline.
+
+---
+
+## Phase 3 — Clean (full pipeline only)
+
+Run the interactive clean command:
+
+```bash
+cd "${CLAUDE_PLUGIN_ROOT}" && node dist/cli.js clean
+```
+
+Or with dry-run:
+```bash
+cd "${CLAUDE_PLUGIN_ROOT}" && node dist/cli.js clean --dry-run
+```
+
+After cleanup, re-run scan to get updated numbers, then show the savings report:
+
+```bash
+cd "${CLAUDE_PLUGIN_ROOT}" && node dist/cli.js report
+```
+
+Present the report box AND the before/after breakdown table to the user.
+
+---
+
+## Phase 4 — Restore
+
+When `/claude-slim restore` is invoked:
+
+```bash
+cd "${CLAUDE_PLUGIN_ROOT}" && node dist/cli.js restore
+```
+
+---
+
+## Language
+
+Detect the user's language from their most recent message. Present all reports, analysis, and explanations in that language. The CLI output is machine-readable — translate only the user-facing interpretation.
+
+## Rules
+
+1. **Never delete.** The CLI moves items to `~/.claude/skills.disabled/`.
+2. **Never modify CLAUDE.md or settings.json.**
+3. **Never disable plugin-managed skills.** Report only.
+4. **Always confirm before executing.** Use `--dry-run` to preview changes.
+5. **If nothing to clean:** respond "Already slim!" and exit.
+6. **Always interpret results.** Never dump raw CLI output without analysis. You are the diagnostic layer — the CLI is the data layer.

package/skills/claude-slim/scripts/scan.sh

@@ -0,0 +1,362 @@
+#!/usr/bin/env bash
+# claude-slim scanner — detects token overhead in Claude Code environments
+set -euo pipefail
+
+OUTPUT_FORMAT="${1:-text}"  # text (default) or json
+SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
+
+# JSON mode: re-run in text mode, pipe through converter
+if [ "$OUTPUT_FORMAT" = "json" ]; then
+  bash "${SCRIPT_DIR}/scan.sh" text | python3 -c "
+import json, sys
+
+data = {}
+for line in sys.stdin:
+    line = line.strip()
+    if not line or line.startswith('==='):
+        continue
+    parts = line.split(':')
+    key = parts[0]
+    if key == 'SKILL':
+        # name may contain ':', size is always the last field
+        data.setdefault('skills', []).append({'name': ':'.join(parts[1:-1]), 'size': int(parts[-1])})
+    elif key == 'BROKEN_SYMLINK':
+        data.setdefault('broken_symlinks', []).append({'name': parts[1], 'target': ':'.join(parts[2:])})
+    elif key == 'LOCAL_SUMMARY':
+        data['local_count'] = int(parts[1])
+        data['local_total_bytes'] = int(parts[2])
+    elif key == 'PLUGIN':
+        p = {'name': parts[1], 'skill_count': int(parts[2])}
+        if len(parts) > 3: p['status'] = parts[3]
+        data.setdefault('plugins', []).append(p)
+    elif key == 'PLUGIN_SKILL':
+        data.setdefault('plugin_skills', []).append({'plugin': parts[1], 'name': ':'.join(parts[2:])})
+    elif key == 'CLAUDE_MD':
+        data['claude_md_bytes'] = int(parts[1])
+    elif key == 'CLAUDE_MD_SECTION':
+        # name may contain ':', size is always last
+        data.setdefault('claude_md_sections', []).append({'name': ':'.join(parts[1:-1]), 'size': int(parts[-1])})
+    elif key == 'MEMORY':
+        # MEMORY:project:name:size — size is always last
+        data.setdefault('memory_files', []).append({
+            'project': ':'.join(parts[1:-2]),
+            'name': parts[-2],
+            'size': int(parts[-1])
+        })
+    elif key == 'MEMORY_SUMMARY':
+        data['memory_count'] = int(parts[1])
+        data['memory_total_bytes'] = int(parts[2])
+    elif key == 'MCP':
+        data['mcp_servers'] = int(parts[1])
+    elif key == 'MCP_SERVER':
+        data.setdefault('mcp_server_list', []).append(':'.join(parts[1:]))
+    elif key == 'ISSUE':
+        issue = {'type': parts[1]}
+        if parts[1] in ('oversized_skill', 'oversized_memory', 'temp_cache', 'stale_project'):
+            # detail is last field, name is everything between type and detail
+            issue['name'] = ':'.join(parts[2:-1])
+            issue['detail'] = parts[-1]
+        elif len(parts) > 3:
+            issue['name'] = parts[2]
+            issue['detail'] = ':'.join(parts[3:])
+        else:
+            issue['name'] = ':'.join(parts[2:])
+        data.setdefault('issues', []).append(issue)
+
+print(json.dumps(data, indent=2))
+"
+  exit 0
+fi
+
+CLAUDE_DIR="${HOME}/.claude"
+SKILLS_DIR="${CLAUDE_DIR}/skills"
+PLUGINS_DIR="${CLAUDE_DIR}/plugins/cache"
+PROJECTS_DIR="${CLAUDE_DIR}/projects"
+
+# Cache `claude plugin list` output once (reused for status + issue detection)
+PLUGIN_LIST_CACHE=""
+if command -v claude >/dev/null 2>&1; then
+  PLUGIN_LIST_CACHE=$(claude plugin list 2>/dev/null || true)
+fi
+
+echo "=== LOCAL SKILLS ==="
+local_count=0
+local_total_bytes=0
+if [ -d "$SKILLS_DIR" ] && [ "$(ls -A "$SKILLS_DIR" 2>/dev/null)" ]; then
+  # Top-level skills
+  for dir in "$SKILLS_DIR"/*/; do
+    [ -d "$dir" ] || continue
+    name=$(basename "$dir")
+    if [ -f "${dir}SKILL.md" ]; then
+      size=$(wc -c < "${dir}SKILL.md" | tr -d ' ')
+      echo "SKILL:${name}:${size}"
+      local_total_bytes=$((local_total_bytes + size))
+      local_count=$((local_count + 1))
+    elif [ -L "${dir}SKILL.md" ] && [ ! -e "${dir}SKILL.md" ]; then
+      echo "BROKEN_SYMLINK:${name}:$(readlink "${dir}SKILL.md" 2>/dev/null || echo unknown)"
+    fi
+  done
+  # Nested skills (e.g., @internal-sys/commit-guide/)
+  for dir in "$SKILLS_DIR"/*/*/; do
+    [ -d "$dir" ] || continue
+    parent=$(basename "$(dirname "$dir")")
+    name="${parent}/$(basename "$dir")"
+    if [ -f "${dir}SKILL.md" ]; then
+      size=$(wc -c < "${dir}SKILL.md" | tr -d ' ')
+      echo "SKILL:${name}:${size}"
+      local_total_bytes=$((local_total_bytes + size))
+      local_count=$((local_count + 1))
+    elif [ -L "${dir}SKILL.md" ] && [ ! -e "${dir}SKILL.md" ]; then
+      echo "BROKEN_SYMLINK:${name}:$(readlink "${dir}SKILL.md" 2>/dev/null || echo unknown)"
+    fi
+  done
+fi
+echo "LOCAL_SUMMARY:${local_count}:${local_total_bytes}"
+
+echo ""
+echo "=== PLUGIN SKILLS ==="
+# Parse disabled plugins from cached output
+disabled_plugins=""
+if [ -n "$PLUGIN_LIST_CACHE" ]; then
+  disabled_plugins=$(echo "$PLUGIN_LIST_CACHE" | python3 -c "
+import sys
+lines = sys.stdin.read().split('\n')
+current_name = None
+for line in lines:
+    line = line.strip()
+    if line.startswith('\u276f'):
+        full = line.split('\u276f')[1].strip()
+        if '@' in full:
+            current_name = full.split('@')[1]
+        else:
+            current_name = full
+    elif 'disabled' in line.lower() and current_name:
+        print(current_name)
+        current_name = None
+    elif 'enabled' in line.lower():
+        current_name = None
+" 2>/dev/null || true)
+fi
+
+if [ -d "$PLUGINS_DIR" ]; then
+  for plugin_dir in "$PLUGINS_DIR"/*/; do
+    [ -d "$plugin_dir" ] || continue
+    pname=$(basename "$plugin_dir")
+    # Skip temp directories
+    [[ "$pname" == temp_local_* ]] && continue
+    # Walk version subdirectories if they exist
+    count=$(find "$plugin_dir" -path "*/skills/*/SKILL.md" 2>/dev/null | wc -l | tr -d ' ')
+    if [ "$count" -gt 0 ]; then
+      # Check if this plugin is disabled
+      status="enabled"
+      echo "$disabled_plugins" | grep -qxF "$pname" 2>/dev/null && status="disabled"
+      echo "PLUGIN:${pname}:${count}:${status}"
+      # Collect skill names for duplicate detection
+      find "$plugin_dir" -path "*/skills/*/SKILL.md" -exec bash -c 'basename "$(dirname "$1")"' _ {} \; 2>/dev/null | while read -r sname; do
+        echo "PLUGIN_SKILL:${pname}:${sname}"
+      done
+    fi
+  done
+fi
+
+echo ""
+echo "=== CLAUDE_MD ==="
+if [ -f "${CLAUDE_DIR}/CLAUDE.md" ]; then
+  size=$(wc -c < "${CLAUDE_DIR}/CLAUDE.md" | tr -d ' ')
+  echo "CLAUDE_MD:${size}"
+  # Section-level breakdown by top-level markdown headers
+  python3 -c "
+import sys
+
+with open(sys.argv[1]) as f:
+    lines = f.readlines()
+
+sections = []
+current_name = None
+current_bytes = 0
+
+for line in lines:
+    if line.startswith('# '):
+        if current_name is not None:
+            sections.append((current_name, current_bytes))
+        elif current_bytes > 0:
+            sections.append(('(preamble)', current_bytes))
+        current_name = line[2:].strip()[:60]
+        current_bytes = len(line.encode('utf-8'))
+    else:
+        current_bytes += len(line.encode('utf-8'))
+
+if current_name is not None:
+    sections.append((current_name, current_bytes))
+
+for name, size in sections:
+    print(f'CLAUDE_MD_SECTION:{name}:{size}')
+" "${CLAUDE_DIR}/CLAUDE.md" 2>/dev/null || true
+else
+  echo "CLAUDE_MD:0"
+fi
+
+echo ""
+echo "=== MEMORY FILES ==="
+mem_total=0
+mem_count=0
+if [ -d "$PROJECTS_DIR" ]; then
+  while IFS= read -r f; do
+    size=$(wc -c < "$f" | tr -d ' ')
+    project=$(echo "$f" | sed "s|${PROJECTS_DIR}/||" | cut -d/ -f1)
+    fname=$(basename "$f")
+    echo "MEMORY:${project}:${fname}:${size}"
+    mem_total=$((mem_total + size))
+    mem_count=$((mem_count + 1))
+  done < <(find "$PROJECTS_DIR" -path "*/memory/*.md" -type f 2>/dev/null)
+fi
+echo "MEMORY_SUMMARY:${mem_count}:${mem_total}"
+
+echo ""
+echo "=== MCP SERVERS ==="
+if [ -f "${CLAUDE_DIR}/settings.json" ]; then
+  python3 -c "
+import json, sys
+with open(sys.argv[1]) as f:
+    d = json.load(f)
+servers = d.get('mcpServers', {})
+print(f'MCP:{len(servers)}')
+for name in sorted(servers.keys()):
+    print(f'MCP_SERVER:{name}')
+" "${CLAUDE_DIR}/settings.json" 2>/dev/null || echo "MCP:0"
+else
+  echo "MCP:0"
+fi
+
+echo ""
+echo "=== ISSUES ==="
+
+# Broken symlinks (scoped to SKILL.md, covers top-level and nested)
+[ -d "$SKILLS_DIR" ] && find "$SKILLS_DIR" -name "SKILL.md" -type l ! -exec test -e {} \; -print 2>/dev/null | while read -r link; do
+  dir=$(dirname "$link")
+  relpath=$(python3 -c "import os,sys; print(os.path.relpath(sys.argv[1], sys.argv[2]))" "$dir" "$SKILLS_DIR" 2>/dev/null || basename "$dir")
+  target=$(readlink "$link" 2>/dev/null || echo "unknown")
+  echo "ISSUE:broken_symlink:${relpath}:${target}"
+done
+
+# Duplicate skills (local name exists in plugin) — top-level + nested
+if [ -d "$SKILLS_DIR" ] && [ -d "$PLUGINS_DIR" ]; then
+  plugin_skill_names=$(find "$PLUGINS_DIR" -path "*/skills/*/SKILL.md" -exec bash -c 'basename "$(dirname "$1")"' _ {} \; 2>/dev/null | sort -u)
+  # Top-level
+  for dir in "$SKILLS_DIR"/*/; do
+    [ -d "$dir" ] || continue
+    name=$(basename "$dir")
+    [ -f "${dir}SKILL.md" ] || continue
+    echo "$plugin_skill_names" | grep -qxF "$name" 2>/dev/null && echo "ISSUE:duplicate:${name}:local+plugin"
+  done
+  # Nested
+  for dir in "$SKILLS_DIR"/*/*/; do
+    [ -d "$dir" ] || continue
+    [ -f "${dir}SKILL.md" ] || continue
+    name=$(basename "$dir")
+    echo "$plugin_skill_names" | grep -qxF "$name" 2>/dev/null && echo "ISSUE:duplicate:$(basename "$(dirname "$dir")")/${name}:local+plugin"
+  done
+fi
+
+# Empty/template skills — top-level + nested
+if [ -d "$SKILLS_DIR" ]; then
+  for dir in "$SKILLS_DIR"/*/; do
+    [ -d "$dir" ] || continue
+    if [ -f "${dir}SKILL.md" ] && grep -q "Replace with description" "${dir}SKILL.md" 2>/dev/null; then
+      echo "ISSUE:template:$(basename "$dir")"
+    fi
+  done
+  for dir in "$SKILLS_DIR"/*/*/; do
+    [ -d "$dir" ] || continue
+    if [ -f "${dir}SKILL.md" ] && grep -q "Replace with description" "${dir}SKILL.md" 2>/dev/null; then
+      echo "ISSUE:template:$(basename "$(dirname "$dir")")/$(basename "$dir")"
+    fi
+  done
+fi
+
+# .skill duplicate directories
+[ -d "$SKILLS_DIR" ] && for dir in "$SKILLS_DIR"/*.skill/; do
+  [ -d "$dir" ] || continue
+  base=$(basename "$dir" .skill)
+  [ -d "${SKILLS_DIR}/${base}" ] && echo "ISSUE:skill_dup:${base}"
+done
+
+# Oversized skill files (>10KB) — top-level + nested
+if [ -d "$SKILLS_DIR" ]; then
+  for dir in "$SKILLS_DIR"/*/; do
+    [ -d "$dir" ] || continue
+    if [ -f "${dir}SKILL.md" ]; then
+      name=$(basename "$dir")
+      size=$(wc -c < "${dir}SKILL.md" | tr -d ' ')
+      [ "$size" -gt 10240 ] && echo "ISSUE:oversized_skill:${name}:${size}"
+    fi
+  done
+  for dir in "$SKILLS_DIR"/*/*/; do
+    [ -d "$dir" ] || continue
+    if [ -f "${dir}SKILL.md" ]; then
+      name="$(basename "$(dirname "$dir")")/$(basename "$dir")"
+      size=$(wc -c < "${dir}SKILL.md" | tr -d ' ')
+      [ "$size" -gt 10240 ] && echo "ISSUE:oversized_skill:${name}:${size}"
+    fi
+  done
+fi
+
+# Oversized memory files (>5KB)
+[ -d "$PROJECTS_DIR" ] && find "$PROJECTS_DIR" -path "*/memory/*.md" -type f -size +5k 2>/dev/null | while read -r f; do
+  size=$(wc -c < "$f" | tr -d ' ')
+  project=$(echo "$f" | sed "s|${PROJECTS_DIR}/||" | cut -d/ -f1)
+  fname=$(basename "$f")
+  echo "ISSUE:oversized_memory:${project}/${fname}:${size}"
+done
+
+# Stale project memory (no files modified in 90+ days)
+if [ -d "$PROJECTS_DIR" ]; then
+  for proj_dir in "$PROJECTS_DIR"/*/; do
+    [ -d "${proj_dir}memory" ] || continue
+    project=$(basename "$proj_dir")
+    # Quick check: any file modified in last 90 days? (maxdepth 1 to match glob below)
+    recent=$(find "${proj_dir}memory" -maxdepth 1 -type f -name "*.md" -mtime -90 2>/dev/null | wc -l | tr -d ' ')
+    if [ "$recent" -eq 0 ]; then
+      # All files older than 90 days — get details via python3
+      detail=$(python3 -c "
+import os, sys, glob, time
+files = glob.glob(sys.argv[1] + '/*.md')
+if files:
+    newest = max(os.path.getmtime(f) for f in files)
+    days = int((time.time() - newest) / 86400)
+    total = sum(os.path.getsize(f) for f in files)
+    print(f'{days}d,{len(files)}files,{total}bytes')
+" "${proj_dir}memory" 2>/dev/null)
+      [ -n "$detail" ] && echo "ISSUE:stale_project:${project}:${detail}"
+    fi
+  done
+fi
+
+# Temp/orphaned plugin cache (failed installs)
+[ -d "$PLUGINS_DIR" ] && for dir in "$PLUGINS_DIR"/temp_local_*/; do
+  [ -d "$dir" ] || continue
+  name=$(basename "$dir")
+  size=$(du -sk "$dir" 2>/dev/null | cut -f1)
+  echo "ISSUE:temp_cache:${name}:${size}KB"
+done
+
+# Disabled plugins (still in cache) — uses cached plugin list
+if [ -n "$PLUGIN_LIST_CACHE" ]; then
+  echo "$PLUGIN_LIST_CACHE" | python3 -c "
+import sys
+lines = sys.stdin.read().split('\n')
+current_name = None
+for line in lines:
+    line = line.strip()
+    if line.startswith('\u276f'):
+        current_name = line.split('\u276f')[1].strip()
+    elif 'disabled' in line.lower() and current_name:
+        print(f'ISSUE:disabled_plugin:{current_name}')
+        current_name = None
+    elif 'enabled' in line.lower():
+        current_name = None
+" 2>/dev/null || true
+fi
+
+echo ""
+echo "=== DONE ==="