Charles Proxy Session Extractor

A Claude Code skill that extracts and analyzes HTTP/HTTPS request and response data from Charles Proxy session files (.chlsj format).

When debugging API integrations or analyzing network traffic, simply mention a .chlsj file and Claude will automatically use this skill to parse and present the data.

Zero dependencies - uses Python standard library only.

Features

Pattern-based filtering - extract requests by URL pattern
Method filtering - filter by HTTP method (GET, POST, PUT, PATCH, DELETE)
Request body inspection - view request bodies for POST/PUT/PATCH operations
JSON export - save filtered responses for further analysis
Summary mode - get quick traffic overview without full response bodies
Pretty-printed output - formatted JSON by default for readability

How It Works

Export a session from Charles Proxy in .chlsj (JSON) format
Reference the file when chatting with Claude Code
Claude automatically detects the file type and uses this skill
Get filtered, formatted request/response data instantly

Prerequisites

Python 3.x (no external dependencies required)
Charles Proxy session files exported in .chlsj format

Installation

For Claude Code Users

Clone this repository into your skills directory:

BASH

# User-level (available in all projects)
git clone https://github.com/yourusername/charles-proxy-extract ~/.claude/skills/charles-proxy-extract

# Project-level (available in specific project)
git clone https://github.com/yourusername/charles-proxy-extract .claude/skills/charles-proxy-extract

That's it! No additional dependencies to install.

Standalone Usage

If you want to use the script without Claude Code, just ensure Python 3.x is installed:

BASH

which python3

Using with Claude Code

Once installed, Claude will automatically suggest this skill when you mention Charles Proxy files. Try phrases like:

"Extract the /api/users responses from session.chlsj"
"Show me POST requests to /logs in this Charles session"
"Analyze the network traffic in debug-session.chlsj"
"Export all /items responses to a JSON file"
"Summarize what's in this Charles session file"
"Show the first /api/auth request"

Claude will use this skill to parse the file and present the results in a clear, formatted way.

Standalone Usage

You can also use the script directly from the command line:

BASH

python3 ./extract_responses.py <file.chlsj> <path_pattern> [options]

Arguments

file - Charles session file (.chlsj format)
pattern - String pattern to match in URL paths (case-sensitive)

Options

Option	Description
`-m, --method METHOD`	Filter by HTTP method (GET, POST, PUT, PATCH, DELETE)
`-f, --first-only`	Show only the first matching request
`-s, --summary-only`	Show summary stats without response bodies
`-o, --output FILE`	Save all responses to a JSON file
`--no-pretty`	Disable pretty-printing of JSON responses
`-h, --help`	Show help message

Note: When filtering by POST, PUT, or PATCH methods, request bodies are automatically displayed along with responses.

Workflow Examples

BASH

# Explore available endpoints
python3 ./extract_responses.py session.chlsj "/" --summary-only

# Inspect specific endpoint
python3 ./extract_responses.py session.chlsj "/logs" --method POST

# Quick peek at first result
python3 ./extract_responses.py session.chlsj "/api/users" --first-only

# Save all responses for analysis
python3 ./extract_responses.py session.chlsj "/items" -o items-response.json

# Filter POST requests with bodies
python3 ./extract_responses.py session.chlsj "/submit" --method POST

Example Output

Summary Mode

Found 145 total requests in session
Pattern '/api/logs' matched 12 requests

Matched paths:
  /api/logs/submit (8 requests)
  /api/logs/query (4 requests)

Methods: POST: 8, GET: 4
Status codes: 200: 11, 404: 1

Full Mode

Request 1/12: POST /api/logs/submit
Status: 200 OK
Timestamp: 2025-12-28T10:15:23Z

Request Body:
{
  "level": "error",
  "message": "Connection timeout"
}

Response:
{
  "id": "log_123",
  "status": "recorded"
}

Export Mode

Creates a JSON file with structure:

JSON

{
  "pattern": "/api/logs",
  "total_requests": 12,
  "extracted_at": "2025-12-28T10:15:23Z",
  "requests": [
    {
      "method": "POST",
      "path": "/api/logs/submit",
      "status": 200,
      "timestamp": "2025-12-28T10:15:23Z",
      "request_body": {...},
      "response": {...}
    }
  ]
}

Tips

Pattern Matching

Patterns use simple substring matching (case-sensitive):

/history matches /history/5fd95c39.../2025-12-04
/items matches both /items/... and /items-by-day/...
/ matches all requests (useful for summary mode)

Best Practices

Start with summary mode - Use --summary-only to understand what's in the session
Narrow down gradually - Start broad, then filter by method or pattern
Use first-only for inspection - Add --first-only when you just need a sample
Export for analysis - Use -o to save data for programmatic analysis

Troubleshooting

"File not found"

Verify the .chlsj file path is correct
Use absolute paths or ensure you're in the right directory

"Invalid JSON"

Ensure the file is a valid Charles Proxy session export
Re-export the session from Charles Proxy in .chlsj format

No matching requests

Pattern matching is case-sensitive - check capitalization
Try broader pattern (e.g., / matches everything)
Use --summary-only to see all available paths

Python not found

Ensure Python 3.x is installed and available in PATH
Try using python instead of python3 or vice versa

Integration Use Cases

This skill is particularly useful for:

Extracting sample data - Generate test fixtures from real API traffic
Debugging integrations - Identify discrepancies between expected and actual API behavior
Documenting APIs - Extract real-world examples for API documentation
Model updates - Find new fields or enum values not yet in your models
Regression testing - Compare API responses before and after changes
Performance analysis - Identify slow endpoints or large payloads

Charles Proxy - HTTP debugging proxy for macOS, Windows, and Linux
Claude Code - AI-powered coding assistant

License

MIT License - see LICENSE.md for details.

Contributing

Contributions are welcome! Feel free to:

Open issues for bugs or feature requests
Submit pull requests for improvements
Share your use cases and feedback