claude-flow-novice 2.18.40 → 2.19.1

package/.claude/skills/{cfn-local-ruvector-accelerator → cfn-codesearch}/AGENT_INTEGRATION_PATTERNS.md

@@ -1,10 +1,10 @@
-# Agent Integration Patterns for AST-Aware RuVector
+# Agent Integration Patterns for AST-Aware CodeSearch
 
-This document describes how agents can integrate with the AST-aware RuVector indexer to achieve sub-50ms query performance for code intelligence tasks.
+This document describes how agents can integrate with the AST-aware CodeSearch indexer to achieve sub-50ms query performance for code intelligence tasks.
 
 ## Overview
 
-The AST-Aware RuVector Accelerator provides:
+The AST-Aware CodeSearch Accelerator provides:
 - Entity-based indexing (functions, structs, traits, etc.)
 - Reference tracking (calls, imports, type usage)
 - Vector embeddings for semantic search
@@ -15,11 +15,11 @@ The AST-Aware RuVector Accelerator provides:
 ### 1. Agent Setup
 
 ```bash
-# Initialize RuVector in the project
-./target/release/local-ruvector init
+# Initialize CodeSearch in the project
+./target/release/local-codesearch init
 
 # Index the codebase
-./target/release/local-ruvector index --path . --types rs,ts,tsx,js,jsx --force
+./target/release/local-codesearch index --path . --types rs,ts,tsx,js,jsx --force
 ```
 
 ### 2. Query Patterns
@@ -28,7 +28,7 @@ The AST-Aware RuVector Accelerator provides:
 
 ```bash
 # Find all functions that use the `Album` type
-./target/release/local-ruvector find --kind function --uses-type Album
+./target/release/local-codesearch find --kind function --uses-type Album
 
 # Example SQL query behind the scenes:
 SELECT e.* FROM entities e
@@ -40,10 +40,10 @@ WHERE e.kind = 'function' AND tu.type_name = 'Album';
 
 ```bash
 # Find all functions that call `create_album`
-./target/release/local-ruvector refs --target create_album --kind calls
+./target/release/local-codesearch refs --target create_album --kind calls
 
 # Exclude calls within the same module
-./target/release/local-ruvector refs --target create_album --kind calls --exclude-module src/album/
+./target/release/local-codesearch refs --target create_album --kind calls --exclude-module src/album/
 
 # Example SQL query:
 SELECT r.* FROM refs r
@@ -56,16 +56,16 @@ WHERE e.name = 'create_album' AND r.ref_kind = 'call'
 
 ```bash
 # Step 1: Find all references to a function
-./target/release/local-ruvector refs --target function_name
+./target/release/local-codesearch refs --target function_name
 
 # Step 2: Find all implementations of a trait
-./target/release/local-ruvector find --kind impl --implements Trait
+./target/release/local-codesearch find --kind impl --implements Trait
 
 # Step 3: Find types used in a module
-./target/release/local-ruvector find --file-path src/module/ --kind struct
+./target/release/local-codesearch find --file-path src/module/ --kind struct
 
 # Step 4: Verify no breaks after refactoring
-./target/release/local-ruvector query "broken reference error" --threshold 0.9
+./target/release/local-codesearch query "broken reference error" --threshold 0.9
 ```
 
 ### 3. Programmatic Integration
@@ -73,8 +73,8 @@ WHERE e.name = 'create_album' AND r.ref_kind = 'call'
 Agents can integrate using the Rust API directly:
 
 ```rust
-use local_ruvector::store_v2::StoreV2;
-use local_ruvector::query_api::QueryEngine;
+use local_codesearch::store_v2::StoreV2;
+use local_codesearch::query_api::QueryEngine;
 
 let store = StoreV2::new(&db_path)?;
 let query_engine = QueryEngine::new(store);
@@ -238,18 +238,18 @@ Based on the current implementation:
 
 ```bash
 # Check database health
-./target/release/local-ruvector stats --detailed
+./target/release/local-codesearch stats --detailed
 
 # Rebuild index
-./target/release/local-ruvector index --force
+./target/release/local-codesearch index --force
 
 # Query optimization
-sqlite3 .ruvector/index.db "EXPLAIN QUERY PLAN SELECT ..."
+sqlite3 .codesearch/index.db "EXPLAIN QUERY PLAN SELECT ..."
 
 # Check embeddings
-./target/release/local-ruvector query "test query" --format json
+./target/release/local-codesearch query "test query" --format json
 ```
 
 ## Conclusion
 
-The AST-Aware RuVector Accelerator provides a powerful foundation for agent-driven code intelligence. By leveraging entity-based indexing and structured queries, agents can achieve sub-50ms response times for complex code analysis tasks.
+The AST-Aware CodeSearch Accelerator provides a powerful foundation for agent-driven code intelligence. By leveraging entity-based indexing and structured queries, agents can achieve sub-50ms response times for complex code analysis tasks.

package/.claude/skills/{cfn-local-ruvector-accelerator → cfn-codesearch}/AGENT_LIFECYCLE_INTEGRATION.md

@@ -1,8 +1,8 @@
-# RuVector Agent Lifecycle Integration
+# CodeSearch Agent Lifecycle Integration
 
 ## Overview
 
-The RuVector Agent Lifecycle Integration automatically captures and indexes agent execution data, enabling semantic search over historical agent decisions, tool usage patterns, and failure recovery strategies.
+The CodeSearch Agent Lifecycle Integration automatically captures and indexes agent execution data, enabling semantic search over historical agent decisions, tool usage patterns, and failure recovery strategies.
 
 ## Architecture
 
@@ -108,7 +108,7 @@ Search historical agent data for guidance on current tasks.
 
 ## Integration with cfn-subagent-stop.sh
 
-The subagent stop hook automatically triggers RuVector ingestion:
+The subagent stop hook automatically triggers CodeSearch ingestion:
 
 ```bash
 # cfn-subagent-stop.sh (excerpt)
@@ -121,8 +121,8 @@ if awk "BEGIN {exit !($AGENT_CONFIDENCE < 0.70)}"; then
     AGENT_SUCCESS="false"
 fi
 
-# Call RuVector ingestion
-./.claude/skills/cfn-local-ruvector-accelerator/ingest-agent-transcript.sh \
+# Call CodeSearch ingestion
+./.claude/skills/cfn-local-codesearch-accelerator/ingest-agent-transcript.sh \
     --transcript "$TRANSCRIPT_ARCHIVE" \
     --agent-id "$AGENT_ID" \
     --agent-type "$AGENT_TYPE" \
@@ -283,9 +283,9 @@ sqlite3 "$DB_PATH" "DELETE FROM agent_failure_patterns WHERE occurrence_count <
 
 ### Log Management
 
-RuVector ingestion logs are written to:
-- `.artifacts/logs/ruvector-ingestion.log` - Transcript ingestion events
-- `.artifacts/logs/ruvector-failure-analysis.log` - Failure pattern analysis
+CodeSearch ingestion logs are written to:
+- `.artifacts/logs/codesearch-ingestion.log` - Transcript ingestion events
+- `.artifacts/logs/codesearch-failure-analysis.log` - Failure pattern analysis
 - `.artifacts/logs/subagent-lifecycle.log` - Hook execution logs
 
 Rotate logs periodically:
@@ -305,14 +305,14 @@ tail -f .artifacts/logs/subagent-lifecycle.log
 
 Verify script exists and is executable:
 ```bash
-ls -l .claude/skills/cfn-local-ruvector-accelerator/ingest-agent-transcript.sh
+ls -l .claude/skills/cfn-local-codesearch-accelerator/ingest-agent-transcript.sh
 ```
 
 ### No patterns found
 
 Run failure analysis manually:
 ```bash
-./.claude/skills/cfn-local-ruvector-accelerator/analyze-agent-failures.sh --days 30
+./.claude/skills/cfn-local-codesearch-accelerator/analyze-agent-failures.sh --days 30
 ```
 
 Check database tables exist:
@@ -335,7 +335,7 @@ Run with broader query:
 
 ## References
 
-- Analysis document: `docs/RUVECTOR_INTEGRATION_ANALYSIS.md`
+- Analysis document: `docs/CODESEARCH_INTEGRATION_ANALYSIS.md`
 - Subagent hooks: `.claude/hooks/cfn-subagent-*.sh`
-- RuVector accelerator: `.claude/skills/cfn-local-ruvector-accelerator/`
+- CodeSearch accelerator: `.claude/skills/cfn-local-codesearch-accelerator/`
 - Lifecycle database: `.claude/skills/cfn-redis-coordination/data/cfn-loop.db`

package/.claude/skills/{cfn-local-ruvector-accelerator → cfn-codesearch}/Cargo.toml

@@ -1,18 +1,23 @@
 [package]
-name = "local-ruvector"
-version = "0.1.0"
+name = "codesearch"
+version = "0.2.0"
 edition = "2021"
 
 [dependencies]
-# Database
+# Database - SQLite for metadata (entities, refs, fast SQL queries)
 rusqlite = { version = "0.31", features = ["bundled", "vtab"] }
+# Database - PostgreSQL/pgvector for embeddings (fast vector search)
+tokio-postgres = { version = "0.7", features = ["with-chrono-0_4", "with-serde_json-1"] }
+deadpool-postgres = { version = "0.12" }
+# pgvector type support for postgres
+pgvector = { version = "0.4", features = ["postgres"] }
 # Vector operations
 ndarray = { version = "0.15", features = ["rayon"], default-features = false }
 # Serialization
 serde = { version = "1.0", features = ["derive"] }
 serde_json = "1.0"
 # CLI
-clap = { version = "4.4", features = ["derive"] }
+clap = { version = "4.4", features = ["derive", "env"] }
 anyhow = "1.0"
 # File system
 walkdir = "2.4"

package/.claude/skills/{cfn-local-ruvector-accelerator → cfn-codesearch}/EXTRACTION_EXAMPLES.md

@@ -1,4 +1,4 @@
-# RuVector Extraction Examples
+# CodeSearch Extraction Examples
 
 This document demonstrates what the TypeScript/JavaScript and text-based fallback extractors capture from different file types.
 
@@ -207,7 +207,7 @@ setup_environment
 
 ## Indexing Results
 
-When you run `local-ruvector index`, the system:
+When you run `local-codesearch index`, the system:
 
 1. **Scans** all files in the project
 2. **Routes** files to appropriate extractors based on extension
@@ -279,9 +279,9 @@ Complete project indexing (10,000 files): ~30-50 seconds depending on file types
 
 After indexing:
 
-1. **Query** using `local-ruvector query`
-2. **Export** results using `local-ruvector export`
-3. **Analyze** dependencies using `local-ruvector refs`
-4. **Monitor** with `local-ruvector stats`
+1. **Query** using `local-codesearch query`
+2. **Export** results using `local-codesearch export`
+3. **Analyze** dependencies using `local-codesearch refs`
+4. **Monitor** with `local-codesearch stats`
 
 See the main README for detailed usage instructions.

package/.claude/skills/{cfn-local-ruvector-accelerator → cfn-codesearch}/IMMEDIATE_ACTION_REQUIRED.md

@@ -1,4 +1,4 @@
-# IMMEDIATE ACTION REQUIRED: RuVector Data Loss Vulnerability
+# IMMEDIATE ACTION REQUIRED: CodeSearch Data Loss Vulnerability
 
 **Date**: 2025-12-11
 **Severity**: 🔴 **CRITICAL**
@@ -8,7 +8,7 @@
 
 ## Executive Summary
 
-The RuVector indexing system has a **critical data loss vulnerability** when operating with a centralized database shared across multiple projects. Reindexing one project can **accidentally delete data from other projects** with identical relative file paths.
+The CodeSearch indexing system has a **critical data loss vulnerability** when operating with a centralized database shared across multiple projects. Reindexing one project can **accidentally delete data from other projects** with identical relative file paths.
 
 **Impact**: ALL projects sharing common file names (e.g., `main.rs`, `lib.rs`, `index.ts`) are at risk.
 
@@ -17,7 +17,7 @@ The RuVector indexing system has a **critical data loss vulnerability** when ope
 ## Vulnerability Details
 
 ### Root Cause
-- Centralized database location: `~/.local/share/ruvector/index_v2.db`
+- Centralized database location: `~/.local/share/codesearch/index_v2.db`
 - File paths stored as **relative paths** (e.g., `src/main.rs`)
 - DELETE query lacks project discriminator: `DELETE FROM entities WHERE file_path = ?`
 
@@ -25,14 +25,14 @@ The RuVector indexing system has a **critical data loss vulnerability** when ope
 
 ```bash
 # Setup
-cd ~/project-a && local-ruvector index src/
+cd ~/project-a && local-codesearch index src/
 # → Stores entities with file_path="src/main.rs"
 
-cd ~/project-b && local-ruvector index src/
+cd ~/project-b && local-codesearch index src/
 # → Stores entities with file_path="src/main.rs"
 
 # Trigger bug
-cd ~/project-a && local-ruvector index --force src/
+cd ~/project-a && local-codesearch index --force src/
 # → DELETE FROM entities WHERE file_path = 'src/main.rs'
 # → Deletes BOTH project-a AND project-b entities!
 ```
@@ -43,7 +43,7 @@ cd ~/project-a && local-ruvector index --force src/
 
 ## Affected Operations
 
-### 1. Reindexing Files (`local-ruvector index --force`)
+### 1. Reindexing Files (`local-codesearch index --force`)
 - **Risk**: Deletes all entities with matching relative path across ALL projects
 - **Detection**: Silent (no errors, no warnings)
 - **Recovery**: Requires full reindex of affected projects
@@ -66,7 +66,7 @@ cd ~/project-a && local-ruvector index --force src/
 # Edit src/paths.rs:
 
 pub fn get_database_path(project_root: &Path) -> Result<PathBuf> {
-    Ok(project_root.join(".ruvector").join("index_v2.db"))
+    Ok(project_root.join(".codesearch").join("index_v2.db"))
 }
 ```
 
@@ -222,7 +222,7 @@ git revert <commit-hash>
 
 # 2. Redeploy previous version
 cargo build --release
-./target/release/local-ruvector --version
+./target/release/local-codesearch --version
 
 # 3. Document issue and replan fix
 ```

package/.claude/skills/{cfn-local-ruvector-accelerator → cfn-codesearch}/README.md

@@ -1,4 +1,4 @@
-# Local RuVector Accelerator
+# Local CodeSearch Accelerator
 
 Fast local semantic code search with multi-project isolation.
 
@@ -6,18 +6,18 @@ Fast local semantic code search with multi-project isolation.
 
 ```bash
 # Initialize
-./target/release/local-ruvector init
+./target/release/local-codesearch init
 
 # Index a project
-./target/release/local-ruvector index --path . --types rs,ts,js
+./target/release/local-codesearch index --path . --types rs,ts,js
 
 # Query patterns
-./target/release/local-ruvector query "authentication middleware"
+./target/release/local-codesearch query "authentication middleware"
 ```
 
 ## Features
 
-- Centralized SQLite database at `~/.local/share/ruvector/index_v2.db`
+- Centralized SQLite database at `~/.local/share/codesearch/index_v2.db`
 - Multi-project isolation with secure path validation
 - OpenAI embeddings for semantic search
 - No Docker or PostgreSQL required
@@ -56,7 +56,7 @@ FOREIGN KEY (entity_id) REFERENCES entities_v2(id) ON DELETE RESTRICT
 
 | Command | Description |
 |---------|-------------|
-| `init` | Initialize RuVector in project |
+| `init` | Initialize CodeSearch in project |
 | `index` | Index code files |
 | `query` | Search patterns semantically |
 | `find` | Structured entity queries |
@@ -68,7 +68,7 @@ FOREIGN KEY (entity_id) REFERENCES entities_v2(id) ON DELETE RESTRICT
 ## Architecture
 
 ```
-~/.local/share/ruvector/
+~/.local/share/codesearch/
   index_v2.db       # Centralized SQLite database
   embeddings/       # Cached embeddings
   cache/            # Query cache

package/.claude/skills/{cfn-local-ruvector-accelerator → cfn-codesearch}/SCHEMA_V2_IMPLEMENTATION.md

@@ -1,4 +1,4 @@
-# Schema v2 Implementation - AST-Aware RuVector Accelerator
+# Schema v2 Implementation - AST-Aware CodeSearch Accelerator
 
 ## Overview
 Successfully transformed flat pattern-based schema (v1) to structured entity/reference schema (v2) for AST-aware code indexing and search.

package/.claude/skills/{cfn-local-ruvector-accelerator → cfn-codesearch}/SECURITY_FINDINGS_SUMMARY.txt

@@ -1,7 +1,7 @@
 SECURITY VALIDATION REPORT - EXECUTIVE SUMMARY
 ==============================================
 
-Project: RuVector V2 Multi-Project Isolation
+Project: CodeSearch V2 Multi-Project Isolation
 Audit Date: 2025-12-11
 Confidence Score: 0.80 (pending critical fixes)
 

package/.claude/skills/{cfn-local-ruvector-accelerator → cfn-codesearch}/SECURITY_IMPLEMENTATION.md

@@ -1,7 +1,7 @@
-# Security Implementation for AST-Aware RuVector Accelerator
+# Security Implementation for AST-Aware CodeSearch Accelerator
 
 ## Overview
-This document describes the comprehensive security controls implemented for the AST-Aware RuVector Accelerator to protect against path traversal attacks, resource exhaustion, and malicious inputs.
+This document describes the comprehensive security controls implemented for the AST-Aware CodeSearch Accelerator to protect against path traversal attacks, resource exhaustion, and malicious inputs.
 
 ## Confidence Score: 0.92
 
@@ -180,6 +180,6 @@ with security_context(base_path, operation) as (validator, monitor):
 
 ## Conclusion
 
-The AST-Aware RuVector Accelerator now includes comprehensive security controls that protect against the most common attack vectors. The implementation follows security best practices and includes thorough testing to ensure effectiveness.
+The AST-Aware CodeSearch Accelerator now includes comprehensive security controls that protect against the most common attack vectors. The implementation follows security best practices and includes thorough testing to ensure effectiveness.
 
 **Confidence Score: 0.92** - High confidence in the security implementation with room for minor improvements in monitoring and alerting.

package/.claude/skills/{cfn-local-ruvector-accelerator → cfn-codesearch}/SECURITY_TESTING_COMPLETION.md

@@ -10,7 +10,7 @@
 
 ## Executive Summary
 
-A production-ready security test suite has been created to validate multi-project isolation in RuVector's centralized SQLite database. The suite consists of **10 comprehensive tests** that collectively detect and prevent the critical cross-project data deletion vulnerability documented in the security audits.
+A production-ready security test suite has been created to validate multi-project isolation in CodeSearch's centralized SQLite database. The suite consists of **10 comprehensive tests** that collectively detect and prevent the critical cross-project data deletion vulnerability documented in the security audits.
 
 **Key Achievement**: All 8 required tests plus 2 bonus integration tests implemented with full documentation and helper infrastructure.
 
@@ -135,7 +135,7 @@ Enables:
 - Path collision scenario: Two projects with same relative filename
 - Impact: All projects sharing common filenames (main.rs, lib.rs, etc.)
 
-**RUVECTOR_ISOLATION_AUDIT.md:**
+**CODESEARCH_ISOLATION_AUDIT.md:**
 - Multi-project attack vectors
 - Path-based isolation weaknesses
 - Transaction and consistency issues
@@ -223,7 +223,7 @@ Tests use:
 
 ### Run All Tests:
 ```bash
-cd /mnt/c/Users/masha/Documents/claude-flow-novice/.claude/skills/cfn-local-ruvector-accelerator
+cd /mnt/c/Users/masha/Documents/claude-flow-novice/.claude/skills/cfn-local-codesearch-accelerator
 cargo test --test test_multi_project_safety --release
 ```
 
@@ -351,7 +351,7 @@ Implementing this test suite provides:
 
 **Total New Code**: 762 lines
 **Total Documentation**: 1400+ lines
-**Absolute Path Prefix**: `/mnt/c/Users/masha/Documents/claude-flow-novice/.claude/skills/cfn-local-ruvector-accelerator/`
+**Absolute Path Prefix**: `/mnt/c/Users/masha/Documents/claude-flow-novice/.claude/skills/cfn-local-codesearch-accelerator/`
 
 ---
 
@@ -379,7 +379,7 @@ Implementing this test suite provides:
 
 ## Conclusion
 
-A comprehensive, production-ready security test suite has been created to validate and enforce multi-project isolation in RuVector. The suite tests all critical vulnerabilities identified in the security audits and provides binary pass/fail validation for security fixes.
+A comprehensive, production-ready security test suite has been created to validate and enforce multi-project isolation in CodeSearch. The suite tests all critical vulnerabilities identified in the security audits and provides binary pass/fail validation for security fixes.
 
 **Status**: Ready for implementation validation
 **Quality**: Production-ready

package/.claude/skills/cfn-codesearch/SKILL.md

@@ -0,0 +1,143 @@
+---
+name: codesearch-code-search
+description: "MANDATORY: Query CodeSearch BEFORE using grep, glob, find, or search. Performs 400x faster semantic and structural code search via SQL on indexed codebase. Use for finding functions, classes, patterns, callers, and implementations. Agents MUST query CodeSearch first; grep is only allowed after CodeSearch returns zero results."
+allowed-tools: Bash, Read, SlashCommand
+---
+
+# CodeSearch Local Semantic Code Search
+
+## MANDATORY: QUERY CODESEARCH BEFORE GREP/GLOB
+
+**This is a PROTOCOL REQUIREMENT, not a suggestion. Failure to query CodeSearch first is a violation.**
+
+### WHY THIS IS MANDATORY
+- CodeSearch SQL: 0.002s | grep: 0.8s (400x slower)
+- Agents using grep first waste tokens and time
+- Index already exists at `~/.local/share/codesearch/index_v2.db`
+
+### ALWAYS USE CODESEARCH FIRST
+```bash
+# Exact name lookup - 0.002s
+sqlite3 ~/.local/share/codesearch/index_v2.db "SELECT file_path, line_number FROM entities WHERE name = 'MyFunction';"
+
+# Fuzzy search - 0.004s
+sqlite3 ~/.local/share/codesearch/index_v2.db "SELECT file_path, line_number FROM entities WHERE name LIKE '%Store%' LIMIT 10;"
+
+# Semantic search
+/codebase-search "authentication middleware pattern"
+```
+
+### GREP IS ONLY ALLOWED WHEN:
+- CodeSearch query returned zero results AND project confirmed not indexed
+- Searching literal strings (error messages, comments, config values)
+- Explicit user request for grep
+
+### FOR CONCEPTUAL QUESTIONS:
+- "Where is X implemented?" → CodeSearch semantic search
+- "Find similar patterns" → CodeSearch embeddings
+- "How is feature Y built?" → CodeSearch first, then read files
+
+## Quick Commands
+
+### Semantic Search (V1 - Embeddings)
+```bash
+# Natural language search
+/codebase-search "authentication middleware pattern"
+/cfn-codesearch-search "error handling in API routes"
+
+# CLI direct (global install preferred)
+local-codesearch query "user login flow" --max-results 5
+```
+
+### Structural Search (V2 - SQL on AST)
+```bash
+# Find all callers of a function
+sqlite3 ~/.local/share/codesearch/index_v2.db \
+  "SELECT * FROM refs WHERE target_name = 'MyFunction';"
+
+# Find all functions in a file
+sqlite3 ~/.local/share/codesearch/index_v2.db \
+  "SELECT name, line_number FROM entities WHERE file_path LIKE '%myfile.rs' AND kind = 'function';"
+
+# Find entities by project (multi-project isolation)
+sqlite3 ~/.local/share/codesearch/index_v2.db \
+  "SELECT COUNT(*) FROM entities WHERE project_root = '/path/to/project';"
+```
+
+## Installation
+
+**Global install (recommended for multi-project use):**
+```bash
+# From claude-flow-novice project
+./scripts/install-codesearch-global.sh
+
+# Verify
+local-codesearch --version
+```
+
+This installs to `~/.local/bin/local-codesearch` for access from any project.
+
+## Prerequisites
+
+**OPENAI_API_KEY is REQUIRED for indexing.** Indexing will fail without a valid key.
+
+```bash
+# Option 1: Export before running
+export OPENAI_API_KEY="sk-..."
+
+# Option 2: Add to shell profile (~/.bashrc or ~/.zshrc)
+echo 'export OPENAI_API_KEY="sk-..."' >> ~/.bashrc
+source ~/.bashrc
+
+# Option 3: Inline with command
+OPENAI_API_KEY="sk-..." ./local-codesearch index --path /project
+```
+
+**Verify key is set:**
+```bash
+echo $OPENAI_API_KEY  # Should show your key (not empty)
+```
+
+## Index Management
+
+```bash
+# Index a project (first time or full rebuild)
+local-codesearch index --path /path/to/project --types rs,ts,py
+
+# Incremental update (after code changes)
+/codebase-reindex
+
+# Check index stats
+sqlite3 ~/.local/share/codesearch/index_v2.db "SELECT project_root, COUNT(*) FROM entities GROUP BY project_root;"
+```
+
+## Key Features
+
+- **Multi-project isolation**: Index multiple projects in single database without data collision
+- **Non-destructive**: Indexing one project never deletes data from other projects
+- **Centralized storage**: `~/.local/share/codesearch/index_v2.db`
+- **Dual search**: V1 semantic (embeddings) + V2 structural (SQL on AST)
+- **Fast**: Rust binary with SQLite backend
+
+## Database Location
+```
+~/.local/share/codesearch/index_v2.db
+```
+
+## For Agents (MANDATORY PROTOCOL)
+
+**DO NOT use grep/glob until you have queried CodeSearch. This is enforced.**
+
+```bash
+# STEP 1: Query CodeSearch FIRST (required)
+/codebase-search "relevant search terms" --top 5
+# Or SQL:
+sqlite3 ~/.local/share/codesearch/index_v2.db "SELECT file_path, line_number FROM entities WHERE name LIKE '%keyword%';"
+
+# STEP 2: Query past errors/patterns
+./.claude/skills/cfn-codesearch/query-agent-patterns.sh "description"
+
+# STEP 3: Only if CodeSearch returns nothing, then use grep
+```
+
+**Violation of this protocol wastes tokens and time. CodeSearch exists to prevent duplicated work.**

package/.claude/skills/{cfn-local-ruvector-accelerator → cfn-codesearch}/analyze-agent-failures.sh

@@ -20,7 +20,7 @@ set -euo pipefail
 SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
 PROJECT_ROOT=$(git rev-parse --show-toplevel 2>/dev/null || echo ".")
 DB_PATH="${PROJECT_ROOT}/.claude/skills/cfn-redis-coordination/data/cfn-loop.db"
-LOG_PATH="${PROJECT_ROOT}/.artifacts/logs/ruvector-failure-analysis.log"
+LOG_PATH="${PROJECT_ROOT}/.artifacts/logs/codesearch-failure-analysis.log"
 
 # Ensure log directory exists
 mkdir -p "$(dirname "$LOG_PATH")"
@@ -50,7 +50,7 @@ while [[ $# -gt 0 ]]; do
     esac
 done
 
-echo "[RuVector] Analyzing agent failures from last $DAYS days" | tee -a "$LOG_PATH"
+echo "[CodeSearch] Analyzing agent failures from last $DAYS days" | tee -a "$LOG_PATH"
 
 # ============================================================================
 # Query Failed Agents
@@ -73,12 +73,12 @@ EOF
 )
 
 if [ -z "$FAILED_AGENTS" ]; then
-    echo "[RuVector] No failures found in last $DAYS days" | tee -a "$LOG_PATH"
+    echo "[CodeSearch] No failures found in last $DAYS days" | tee -a "$LOG_PATH"
     exit 0
 fi
 
 FAILURE_COUNT=$(echo "$FAILED_AGENTS" | wc -l)
-echo "[RuVector] Found $FAILURE_COUNT failed agent executions" | tee -a "$LOG_PATH"
+echo "[CodeSearch] Found $FAILURE_COUNT failed agent executions" | tee -a "$LOG_PATH"
 
 # ============================================================================
 # Extract Failure Patterns
@@ -150,7 +150,7 @@ SET
 WHERE id = '$EXISTING_ID';
 SQL_EOF
 
-                echo "[RuVector] Updated pattern: $EXISTING_ID (count: $NEW_COUNT)" | tee -a "$LOG_PATH"
+                echo "[CodeSearch] Updated pattern: $EXISTING_ID (count: $NEW_COUNT)" | tee -a "$LOG_PATH"
 
             else
                 # Create new pattern
@@ -177,13 +177,13 @@ VALUES (
 SQL_EOF
 
                 PATTERNS_CREATED=$((PATTERNS_CREATED + 1))
-                echo "[RuVector] Created pattern: $PATTERN_ID ($AGENT_TYPE / $FAILURE_MODE)" | tee -a "$LOG_PATH"
+                echo "[CodeSearch] Created pattern: $PATTERN_ID ($AGENT_TYPE / $FAILURE_MODE)" | tee -a "$LOG_PATH"
             fi
         fi
     fi
 done
 
-echo "[RuVector] Created $PATTERNS_CREATED new failure patterns" | tee -a "$LOG_PATH"
+echo "[CodeSearch] Created $PATTERNS_CREATED new failure patterns" | tee -a "$LOG_PATH"
 
 # ============================================================================
 # Identify High-Impact Patterns
@@ -205,7 +205,7 @@ EOF
 
 if [ -n "$HIGH_IMPACT" ]; then
     echo "" | tee -a "$LOG_PATH"
-    echo "[RuVector] High-impact failure patterns (>= $MIN_OCCURRENCES occurrences):" | tee -a "$LOG_PATH"
+    echo "[CodeSearch] High-impact failure patterns (>= $MIN_OCCURRENCES occurrences):" | tee -a "$LOG_PATH"
     echo "$HIGH_IMPACT" | while IFS='|' read -r TYPE MODE COUNT CAUSE; do
         echo "  - $TYPE / $MODE: $COUNT occurrences" | tee -a "$LOG_PATH"
         echo "    Cause: $(echo "$CAUSE" | head -c 80)..." | tee -a "$LOG_PATH"
@@ -220,7 +220,7 @@ fi
 # and extract their approaches as recovery strategies
 
 echo "" | tee -a "$LOG_PATH"
-echo "[RuVector] Searching for recovery strategies..." | tee -a "$LOG_PATH"
+echo "[CodeSearch] Searching for recovery strategies..." | tee -a "$LOG_PATH"
 
 RECOVERIES_FOUND=0
 
@@ -259,14 +259,14 @@ SQL_EOF
     fi
 done
 
-echo "[RuVector] Identified $RECOVERIES_FOUND recovery strategies" | tee -a "$LOG_PATH"
+echo "[CodeSearch] Identified $RECOVERIES_FOUND recovery strategies" | tee -a "$LOG_PATH"
 
 # ============================================================================
 # Summary Report
 # ============================================================================
 
 echo "" | tee -a "$LOG_PATH"
-echo "[RuVector] Failure Analysis Summary:" | tee -a "$LOG_PATH"
+echo "[CodeSearch] Failure Analysis Summary:" | tee -a "$LOG_PATH"
 echo "  - Analyzed: $FAILURE_COUNT failed agents" | tee -a "$LOG_PATH"
 echo "  - New patterns: $PATTERNS_CREATED" | tee -a "$LOG_PATH"
 echo "  - Recovery strategies: $RECOVERIES_FOUND" | tee -a "$LOG_PATH"