메뉴
⚡ AUTO-INVOKE when user asks: 'audit', 'investigate', 'how does X work', 'find all', 'where is', 'trace', 'understand', 'map the codebase', 'comprehensive'. MUST run BEFORE Read/Glob when planning to read 3+ files. Prevents tool familiarity bias toward native tools.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
| name | code-search-selector |
| description | ⚡ AUTO-INVOKE when user asks: 'audit', 'investigate', 'how does X work', 'find all', 'where is', 'trace', 'understand', 'map the codebase', 'comprehensive'. MUST run BEFORE Read/Glob when planning to read 3+ files. Prevents tool familiarity bias toward native tools. |
| allowed-tools | Bash, Read, AskUserQuestion |
╔══════════════════════════════════════════════════════════════════════════════╗
║ ║
║ ⚡ THIS SKILL AUTO-TRIGGERS ON THESE KEYWORDS: ║
║ ║
║ "audit" | "investigate" | "how does X work" | "find all" | "where is" ║
║ "trace" | "understand" | "map the codebase" | "comprehensive" ║
║ "all integration points" | "find implementations" | "architecture" ║
║ ║
║ 🚫 INTERCEPTION: Triggers when about to Read 3+ files OR Glob broadly ║
║ ║
╚══════════════════════════════════════════════════════════════════════════════╝
The Tool Familiarity Bias Problem:
You have "native" tools (Read, Glob, Grep) that are always available with predictable output. These feel safe. But they produce INFERIOR results for semantic queries.
The "Known File Path" Trap:
When a prompt mentions specific file paths, your instinct is to Read directly. RESIST THIS. Semantic search provides CONTEXT around those files that direct reads miss.
The Parallelization Excuse:
"Let me Read files while agents work" is inefficient. Claudemem's indexed data is FASTER and provides better context.
This skill ensures you use the RIGHT tool for code search tasks. Using Grep when claudemem is indexed is a critical mistake that produces inferior results.
❌ WRONG: User asks "How does authentication work?"
→ You use: grep -r "auth" src/
→ Result: 500 lines of noise, no understanding
✅ RIGHT: User asks "How does authentication work?"
→ You check: claudemem status
→ You use: claudemem search "authentication login flow JWT"
→ Result: Top 10 semantically relevant code chunks
┌─────────────────────────────────────────────────────────────────┐
│ WHAT IS THE USER ASKING? │
├─────────────────────────────────────────────────────────────────┤
│ │
│ "Find all X" → SEMANTIC (go to Step 2) │
│ "How does X work" → SEMANTIC (go to Step 2) │
│ "Audit X integration" → SEMANTIC (go to Step 2) │
│ "Map the data flow" → SEMANTIC (go to Step 2) │
│ "Understand architecture" → SEMANTIC (go to Step 2) │
│ "Trace X through code" → SEMANTIC (go to Step 2) │
│ "Find implementations" → SEMANTIC (go to Step 2) │
│ "What patterns are used" → SEMANTIC (go to Step 2) │
│ │
│ "Find exact string 'foo'" → EXACT MATCH (use Grep, skip tree) │
│ "Count occurrences of X" → EXACT MATCH (use Grep, skip tree) │
│ "Find symbol UserService" → EXACT MATCH (use Grep, skip tree) │
│ │
└─────────────────────────────────────────────────────────────────┘
# ALWAYS run this before semantic search
claudemem status
Interpret the output:
| Status | What It Means | Next Action |
|---|---|---|
| Shows chunk count (e.g., "938 chunks") | ✅ Indexed | USE CLAUDEMEM (Step 3) |
| "No index found" | ❌ Not indexed | Offer to index (Step 2b) |
| "command not found" | ❌ Not installed | Fall back to Detective agent |
AskUserQuestion({
questions: [{
question: "Claudemem is not indexed. Index now for better semantic search results?",
header: "Index?",
multiSelect: false,
options: [
{ label: "Yes, index now (Recommended)", description: "Takes 1-2 minutes, enables semantic search" },
{ label: "No, use grep instead", description: "Faster but less accurate for semantic queries" }
]
}]
})
If user says yes:
claudemem index -y
IF CLAUDEMEM IS INDEXED (from Step 2):
# Get role-specific guidance first
claudemem ai developer # or architect, tester, debugger
# Then search semantically
claudemem search "authentication login JWT token validation" -n 15
IF CLAUDEMEM IS NOT AVAILABLE:
Use the detective agent:
Task({
subagent_type: "code-analysis:detective",
description: "Investigate [topic]",
prompt: "Use semantic search to find..."
})
╔══════════════════════════════════════════════════════════════════╗
║ ❌ FORBIDDEN when claudemem is indexed: ║
║ ║
║ grep -r "pattern" src/ # Use claudemem search instead ║
║ Grep tool for semantic queries # Use claudemem search instead ║
║ Glob to find implementations # Use claudemem search instead ║
║ find . -name "*.ts" | xargs... # Use claudemem search instead ║
║ ║
║ These tools are for EXACT MATCHES only, not semantic search. ║
╚══════════════════════════════════════════════════════════════════╝
| User Request | ❌ DON'T Use | ✅ DO Use |
|---|---|---|
| "Audit all API endpoints" | grep -r "router|endpoint" | claudemem search "API endpoint route handler" |
| "How does auth work?" | grep -r "auth|login" | claudemem search "authentication login flow" |
| "Find all database queries" | grep -r "prisma|query" | claudemem search "database query SQL prisma" |
| "Map the data flow" | grep -r "transform|map" | claudemem search "data transformation pipeline" |
| "What's the architecture?" | ls -la src/ | claudemem search "architecture layer service" |
| "Find error handling" | grep -r "catch|error" | claudemem search "error handling exception" |
| "Trace user creation" | grep -r "createUser" | claudemem search "user creation registration" |
✅ Use Grep for:
grep -r "DEPRECATED_FLAG" src/grep -c "import React" src/**/*.tsxgrep -r "class UserService" src/grep -r "TODO:\|FIXME:" src/❌ Never use Grep for:
After using this skill's decision tree, invoke the appropriate detective:
| Investigation Type | Detective Skill |
|---|---|
| Architecture patterns | code-analysis:architect-detective |
| Implementation details | code-analysis:developer-detective |
| Test coverage | code-analysis:tester-detective |
| Bug root cause | code-analysis:debugger-detective |
| Comprehensive audit | code-analysis:ultrathink-detective |
┌─────────────────────────────────────────────────────────────────┐
│ CODE SEARCH QUICK REFERENCE │
├─────────────────────────────────────────────────────────────────┤
│ │
│ 1. ALWAYS check first: claudemem status │
│ │
│ 2. If indexed: claudemem search "semantic query" │
│ │
│ 3. For exact matches: Grep tool (only this case!) │
│ │
│ 4. For deep analysis: Task(code-analysis:detective) │
│ │
│ ⚠️ GREP IS FOR EXACT MATCHES, NOT SEMANTIC UNDERSTANDING │
│ │
└─────────────────────────────────────────────────────────────────┘
Before ANY code investigation task, verify:
claudemem status to check index╔══════════════════════════════════════════════════════════════════════════════╗
║ STOP BEFORE BULK FILE OPERATIONS ║
╠══════════════════════════════════════════════════════════════════════════════╣
║ ║
║ INTERCEPT TRIGGER: Before executing any of these: ║
║ ║
║ • Read 3+ files in same directory ║
║ • Glob with broad patterns (**/*.ts, **/*.py) ║
║ • Sequential reads to "understand" a feature ║
║ • "Let me read files while agents work" ║
║ ║
║ ASK YOURSELF: ║
║ 1. Is claudemem indexed? (claudemem status) │
║ 2. Can this be ONE semantic query instead of N file reads? ║
║ 3. Am I falling into tool familiarity bias? ║
║ ║
╚══════════════════════════════════════════════════════════════════════════════╝
❌ About to do:
Read src/services/auth/login.ts
Read src/services/auth/session.ts
Read src/services/auth/jwt.ts
Read src/services/auth/middleware.ts
Read src/services/auth/types.ts
Read src/services/auth/utils.ts
✅ Do instead:
claudemem search "authentication login session JWT middleware" -n 15
❌ About to do:
Glob pattern: src/services/prime/**/*.ts
Then read all 12 matches sequentially
✅ Do instead:
claudemem search "Prime API integration service endpoints" -n 20
❌ Parallelization trap:
"Let me Read these 5 files while the detective agent works..."
✅ Do instead:
Trust the detective agent to use claudemem.
Don't duplicate work with inferior Read/Glob.
| Anti-Pattern | Why It's Wrong | Correct Alternative |
|---|---|---|
| Reading 5+ files sequentially | Token waste, no ranking | claudemem search once |
| Glob → Read all matches | No semantic understanding | claudemem search with concept |
| "Files mentioned, let me Read" | Misses context around files | Search semantically first |
| Grep for "how does X work" | Text match ≠ meaning | claudemem search |
| Read while agents work | Duplicate inferior work | Trust agent's claudemem usage |
┌─────────────────────────────────────────────────────────────────┐
│ CORRECT INVESTIGATION FLOW │
├─────────────────────────────────────────────────────────────────┤
│ │
│ 1. TASK ARRIVES with keywords: │
│ "audit", "investigate", "how does", "find all", etc. │
│ │
│ 2. AUTO-TRIGGER this skill (code-search-selector) │
│ │
│ 3. CHECK: claudemem status │
│ • If indexed → Use claudemem search │
│ • If not → Index first OR launch detective agent │
│ │
│ 4. SEARCH SEMANTICALLY: │
│ claudemem search "concept query" -n 15 │
│ │
│ 5. ONLY THEN Read specific files/lines from results │
│ │
│ ⚠️ NEVER start with Read/Glob for semantic tasks │
│ │
└─────────────────────────────────────────────────────────────────┘
Maintained by: MadAppGang Plugin: code-analysis v2.2.0 Purpose: Prevent tool familiarity bias, intercept multi-file reads, enforce semantic search
CRITICAL - Guide for using Claudish CLI ONLY through sub-agents to run Claude Code with OpenRouter models (Grok, GPT-5, Gemini, MiniMax). NEVER run Claudish directly in main context unless user explicitly requests it. Use when user mentions external AI models, Claudish, OpenRouter, or alternative models. Includes mandatory sub-agent delegation patterns, agent selection guide, file-based instructions, and strict rules to prevent context window pollution.
Multi-agent code analysis orchestration using claudemem. Share claudemem output across parallel agents. Enables parallel investigation, consensus analysis, and role-based command mapping.
Integration guide for using detective skills across plugins. Maps agent roles from frontend, bun, and other plugins to appropriate detective skills. Developer agents should use developer-detective, architect agents should use architect-detective, etc.
⚡ PRIMARY TOOL for: 'why is X broken', 'find bug source', 'root cause analysis', 'trace error', 'debug issue', 'find where X fails'. Uses claudemem v0.3.0 AST with context command for call chain analysis. GREP/FIND/GLOB ARE FORBIDDEN.
⛔ INTERCEPT TRIGGER: Automatically invoked BEFORE Read 3+ files OR Glob with broad patterns. Validates whether bulk file operations should be replaced with semantic search. Prevents token waste from sequential file reads.
⚡ PRIMARY TOOL for: 'what's tested', 'find test coverage', 'audit test quality', 'missing tests', 'edge cases', 'test patterns'. Uses claudemem v0.3.0 AST with callers analysis for test discovery. GREP/FIND/GLOB ARE FORBIDDEN.