file-search
The file-search skill provides fast text and structural code pattern searching using ripgrep for regex-based text matching and ast-grep for syntax-aware code analysis. Use this skill when understanding codebases, finding function or class usages before refactoring, locating specific code patterns like error handling or security issues, analyzing dependencies, or searching for documentation markers like TODOs and FIXMEs across large projects.
git clone --depth 1 https://github.com/massgen/MassGen /tmp/file-search && cp -r /tmp/file-search/massgen/skills/file-search ~/.claude/skills/file-searchSKILL.md
# File Search Skill
Search code efficiently using ripgrep for text patterns and ast-grep for structural code patterns.
## Purpose
The file-search skill provides access to two powerful search tools pre-installed in MassGen environments:
1. **ripgrep (rg)**: Ultra-fast text search with regex support for finding strings, patterns, and text matches
2. **ast-grep (sg)**: Syntax-aware structural search for finding code patterns based on abstract syntax trees
Use these tools to understand codebases, find usage patterns, analyze impact of changes, and locate specific code constructs. Both tools are significantly faster than traditional `grep` or `find` commands.
## When to Use This Skill
Use the file-search skill when:
- Understanding a new codebase (finding entry points, key classes)
- Finding all usages of a function, class, or variable before refactoring
- Locating specific code patterns (error handling, API calls, etc.)
- Searching for security issues (hardcoded credentials, SQL queries, eval usage)
- Analyzing dependencies and imports
- Finding TODOs, FIXMEs, or code comments
Choose **ripgrep** for:
- Text-based searches (strings, comments, variable names)
- Fast, simple pattern matching across many files
- When the exact code structure doesn't matter
Choose **ast-grep** for:
- Structural code searches (function signatures, class definitions)
- Syntax-aware matching (understanding code semantics)
- Complex refactoring (finding specific code patterns)
## Invoking Search Tools
In MassGen, use the `execute_command` tool to run ripgrep and ast-grep:
```python
# Using ripgrep
execute_command("rg 'pattern' --type py src/")
# Using ast-grep
execute_command("sg --pattern 'class $NAME { $$$ }' --lang python")
```
Both tools are pre-installed in MassGen Docker containers and available via shell execution.
## Targeting Your Searches
**CRITICAL**: Always start with targeted, narrow searches to avoid overwhelming results. Getting thousands of matches makes analysis impossible and wastes tokens.
These strategies apply to **both ripgrep and ast-grep**.
### Scope-Limiting Strategies
Apply these strategies from the start to target searches effectively:
1. **Specify File Types/Languages**: Always filter by language
```bash
# Ripgrep
rg "function" --type py --type js
# AST-grep
sg --pattern 'function $NAME($$$) { $$$ }' --lang js
```
2. **Target Specific Directories**: Search in likely locations first
```bash
# Ripgrep
rg "LoginService" src/services/
# AST-grep
sg --pattern 'class LoginService { $$$ }' src/services/
```
3. **Use Specific Patterns**: Make patterns as specific as possible
```bash
# Ripgrep: BAD - too broad
rg "user"
# Ripgrep: GOOD - more specific
rg "class.*User.*Service" --type py
# AST-grep: BAD - too broad
sg --pattern '$X'
# AST-grep: GOOD - more specific
sg --pattern 'class $NAME extends UserService { $$$ }' --lang js
```
4. **Limit Result Count**: Use head to cap results
```bash
# Ripgrep
rg "import" --type py | head -20
rg "TODO" --count
# AST-grep
sg --pattern 'import $X from $Y' --lang js | head -20
```
### Progressive Search Refinement
When exploring unfamiliar code, use this workflow:
**Ripgrep example:**
```bash
# Step 1: Count matches to assess scope
rg "pattern" --count --type py
# Step 2: If too many results, add more filters
rg "pattern" --type py src/ --glob '!tests'
# Step 3: Show limited results to inspect
rg "pattern" --type py src/ | head -30
# Step 4: Once confirmed, get full results or target further
rg "pattern" --type py src/specific_module/
```
**AST-grep example:**
```bash
# Step 1: Assess scope with broad structural pattern
sg --pattern 'function $NAME($$$) { $$$ }' --lang js | head -10
# Step 2: If too many results, narrow to specific directory
sg --pattern 'function $NAME($$$) { $$$ }' --lang js src/
# Step 3: Make pattern more specific
sg --pattern 'async function $NAME($$$) { $$$ }' --lang js src/
# Step 4: Target exact location
sg --pattern 'async function $NAME($$$) { $$$ }' --lang js src/services/
```
### When You Get Too Many Results
If a search returns hundreds of matches (applies to both `rg` and `sg`):
1. **Add file type/language filters**: `--type py` (rg) or `--lang python` (sg)
2. **Narrow directory scope**: Search `src/` instead of `.`
3. **Make pattern more specific**: Add context around the pattern
4. **Use word boundaries**: `-w` flag for whole words only (rg)
5. **Pipe to head**: Limit output with `| head -50`
6. **Exclude test files**: `--glob '!*test*'` (rg) or avoid test directories (sg)
**Example of refinement:**
```bash
# Step 1: Too broad (10,000+ matches)
rg "error"
# Step 2: Add file type (1,000 matches)
rg "error" --type py
# Step 3: Add directory scope (200 matches)
rg "error" --type py src/
# Step 4: Make pattern specific (20 matches)
rg "raise.*Error" --type py src/
# Step 5: Target exact location (5 matches)
rg "raise.*Error" --type py src/services/
```
## How to Use
### Ripgrep (rg)
```bash
# Basic text search
rg "pattern" --type py --type js
# Common flags
-i # Case-insensitive
-w # Match whole words only
-l # Show only filenames
-n # Show line numbers
-C 3 # Show 3 lines of context
--count # Count matches per file
--glob '!dir' # Exclude directory
# Examples
rg "function.*login" --type js src/
rg -i "TODO" --count
rg "auth|login|session" --type py
```
### AST-Grep (sg)
```bash
# Structural code search
sg --pattern 'function $NAME($$$) { $$$ }' --lang js
# Metavariables
$VAR # Matches single AST node
$$$ # Matches zero or more nodes
# Examples
sg --pattern 'class $NAME { $$$ }' --lang python
sg --pattern 'import $X from $Y' --lang js
sg --pattern 'async function $NAME($$$) { $$$ }' src/
```
## Common Search Patterns
```bash
# Security issues
rg -i "password\s*=\s*['\"]" --type py
rg "\beval\(" --type js
# TODGuide to audio generation and understanding in MassGen. Covers text-to-speech, music, sound effects, and audio understanding across ElevenLabs and OpenAI backends.
Complete guide for integrating a new LLM backend into MassGen. Use when adding a new provider (e.g., Codex, Mistral, DeepSeek) or when auditing an existing backend for missing integration points. Covers all ~15 files that need touching.
Guide for creating evolving skills - detailed workflow plans that capture what you'll do, what tools you'll create, and learnings from execution. Use this when starting a new task that could benefit from a reusable workflow.
Guide to image generation and editing in MassGen. Use when creating images, editing existing images, iterating on image designs, or choosing between image backends (OpenAI, Google Gemini/Imagen, Grok, OpenRouter).
Guide for creating properly structured YAML configuration files for MassGen. This skill should be used when agents need to create new configs for examples, case studies, testing, or demonstrating features.
Guide for using MassGen to develop and improve itself. This skill should be used when agents need to run MassGen experiments programmatically (using automation mode) OR analyze terminal UI/UX quality (using visual evaluation tools). These are mutually exclusive workflows for different improvement goals.
Run MassGen experiments and analyze logs using automation mode, logfire tracing, and SQL queries. Use this skill for performance analysis, debugging agent behavior, evaluating coordination patterns, and improving the logging structure, or whenever an ANALYSIS_REPORT.md is needed in a log directory.
Guide for following MassGen's release documentation workflow. This skill should be used when preparing release documentation, updating changelogs, writing case studies, or maintaining project documentation across releases.