har-mcp/
├── har_mcp/
│   ├── __init__.py
│   ├── __main__.py        # python -m har_mcp entry point
│   ├── server.py           # MCP server — tool registration & wiring
│   ├── har_parser.py       # HAR 1.2 parsing, decompression, URL normalization
│   ├── analyzer.py         # Smart analysis (auth, patterns, chains, comparison)
│   ├── models.py           # Data models for HAR structures
│   ├── security.py         # Path validation, secret redaction, content sanitization
│   ├── storage.py          # In-memory storage backend
│   └── config.py           # Environment variable configuration
├── tests/
│   ├── conftest.py
│   ├── test_parser.py
│   ├── test_security.py
│   ├── test_tools.py
│   └── fixtures/
│       └── sample.har      # Fake sample HAR for testing
├── sample.har              # Fake sample HAR (example.com only)
├── setup.py
├── requirements.txt
└── README.md

# Clone the repo
git clone https://github.com/user/har-mcp.git
cd har-mcp

# Install in editable mode (recommended)
pip install -e .

# Or install dependencies only
pip install -r requirements.txt

{
  "mcpServers": {
    "har-analysis": {
      "command": "python",
      "args": ["-m", "har_mcp.server"],
      "env": {
        "HAR_ALLOWED_ROOT": "/path/to/your/har-files"
      }
    }
  }
}

{
  "mcpServers": {
    "har-analysis": {
      "command": "C:\\Users\\you\\AppData\\Local\\Programs\\Python\\Python313\\python.exe",
      "args": ["-m", "har_mcp.server"],
      "env": {
        "HAR_ALLOWED_ROOT": "C:\\path\\to\\har-files"
      }
    }
  }
}

claude mcp add har-analysis -- python -m har_mcp.server

{
  "mcpServers": {
    "har-analysis": {
      "command": "python",
      "args": ["-m", "har_mcp.server"]
    }
  }
}

1. Load the HAR file
   → load_har("recording.har")
   → Returns: har_id, version, request_count, domains

2. Get an overview
   → get_statistics(har_id)
   → See methods, status codes, avg response time, largest responses

3. Find all API endpoints
   → list_endpoints(har_id)
   → GET /api/users (12x, [200])
   → POST /api/login (3x, [200, 401])

4. Search for authentication
   → search_requests(har_id, query="token", scope="all")
   → Find where tokens appear in headers, bodies, cookies

5. Trace a specific token
   → trace_value(har_id, value="eyJ...")
   → See it first appear in POST /login response, then used in subsequent requests

6. Analyze the auth flow
   → analyze_auth(har_id)
   → Detects: Bearer token, session cookie, CSRF token

7. Compare two similar requests
   → compare_requests(har_id, request_a=5, request_b=8)
   → See what changed (different query params, different response)

8. Export as cURL
   → export_request_as_curl(har_id, request_id=12)
   → Ready-to-run command with secrets redacted

# Install test dependencies
pip install -e ".[dev]"

# Run all tests (111 tests)
pytest tests/ -v

# Run specific test file
pytest tests/test_parser.py -v
pytest tests/test_security.py -v
pytest tests/test_tools.py -v

# Run with coverage
pytest tests/ --cov=har_mcp