ASL Parser

A powerful CLI tool to parse and analyze codebases, generating comprehensive agent specifications in agents.md.

Installation

Global Installation (Recommended)

npm install -g asl-parser

Use the asl command from anywhere:

asl parse ./

Local Installation

npm install asl-parser

Use via npx:

npx asl parse ./

Quick Start

Parse Command

Parse the current directory:

asl parse

Or specify a custom path:

asl parse ./my-project

Options:

| Option | Description | Default | |--------|-------------|---------| | -o, --output <file> | Output file name | agents.md |

Example with custom output:

asl parse ./ -o my-agents.md

Prompt Command

Analyze a prompt to identify impacted files, functions, and lines. Results print to terminal only—no file created.

asl prompt "Add authentication to the login function"

With custom path:

asl prompt "Update the API endpoint handler" ./my-project

RAG Search Search codebase using natural language:

asl rag "your search query"

Examples: bash

asl rag "database connection" -p ./src
asl rag "error handling" -k 20

Options:

-p, --path – Codebase path (default: .)
-k, --topk – Number of results (default: 10)

Returns: Relevant code sections with line numbers and relevance scores.

What It Does:

• Extracts keywords, function names, file paths, and tech terms from your prompt
• Scans the codebase for relevant files, functions, and classes
• Prints a focused impact report to the terminal

Output Includes:

• Files affected and relevance scores
• Functions/classes impacted
• Specific line numbers (e.g., L42, L87)
• File metadata (extension, line count)

Benefits:

• Save tokens by analyzing only relevant code sections
• Quickly identify what needs to be changed
• See impacted functions and classes at a glance
• Get specific line numbers for pinpoint edits

Output

The parse command generates an agents.md file containing:

• Summary Statistics – Total files, lines of code, and file sizes
• Files by Extension – Breakdown of file types in your codebase
• File Analysis – Detailed breakdown for each file:
  • File paths and sizes
  • Code vs. comment line counts
  • Detected classes and functions
• Agent Specifications – Structured JSON data ready for AI agents

Features

✨ Multi-Language Support – JavaScript, TypeScript, Python, Java, C++, Go, Rust, and more

🔍 Smart Detection – Automatically detects functions and classes in code files

📊 Comprehensive Docs – Generates markdown documentation and structured JSON output

⚡ Impact Analysis – Prompt-based impact analysis to identify affected files and lines without full codebase scans

🚀 Auto-Ignored Directories – Skips common directories like node_modules, .git, dist, and more 🔍 RAG Search – Natural language code search without API keys

Supported Languages

| Language | Extensions | |----------|-----------| | JavaScript/TypeScript | .js, .jsx, .ts, .tsx | | Python | .py | | Java | .java | | C/C++ | .c, .cpp, .h | | Go | .go | | Rust | .rs | | Ruby | .rb | | PHP | .php | | Swift | .swift | | Kotlin | .kt | | And more... | — |

Ignored Directories

The following directories are automatically skipped during analysis:

• node_modules, vendor
• .git, .vscode, .idea
• .next, .nuxt
• dist, build, target
• .cache, coverage
• __pycache__, bin, obj

Example

asl rag "how does rag works"
>>

🔍 RAG Search: "how does rag works"

📚 Indexing codebase...
✅ Indexed 114 code sections

✅ Found 10 relevant sections:

📍 Query: "how does rag works"

1. 📄 lib\commands\rag-command.js
   Lines 181-200 | Relevance: 30%
         process.exit(1);
... }

2. 📄 lib\rag\minimal-rag.js
   Lines 1-20 | Relevance: 22%
   // src/rag/minimal-rag.js
...Drop-in RAG module for asl-parser

3. 📄 lib\commands\rag-command.js
   Lines 1-20 | Relevance: 15%
   // lib/rag-command.js
...Simple RAG search command for asl-parser

4. 📄 lib\commands\rag-command.js
   Lines 166-185 | Relevance: 15%
         console.log(chalk.white(`   ${snippet}...`));
      console.log...

5. 📄 lib\commands\rag-command.js
   Lines 16-35 | Relevance: 14%

...st CODE_EXTENSIONS = [

6. 📄 lib\commands\rag-command.js
   Lines 196-204 | Relevance: 11%
       rag.displayResults(results, query);
...

7. 📄 lib\rag\minimal-rag.js
   Lines 76-95 | Relevance: 7%
     }
...

8. 📄 lib\promptAnalyzer.js
   Lines 31-50 | Relevance: 4%
     });
...

9. 📄 lib\promptAnalyzer.js
   Lines 91-110 | Relevance: 4%
       if (filePath.includes(term) || fileName.includes(term)) {       
      s...

10. 📄 lib\promptAnalyzer.js
   Lines 106-125 | Relevance: 4%
   /**
...Score how relevant a function/class is to the prompt