generate_filesystem.py

#!/usr/bin/env python3
"""
Generate filesystem.json manifest from pentest artifacts directory
"""

import json
import os
from pathlib import Path

def read_file_safe(filepath):
    """Safely read file contents, return None if binary or unreadable"""
    try:
        with open(filepath, 'r', encoding='utf-8') as f:
            return f.read()
    except (UnicodeDecodeError, PermissionError):
        return "[Binary file or unreadable]"

def build_tree(directory):
    """Recursively build directory tree with file contents"""
    tree = {
        "name": directory.name,
        "type": "directory",
        "children": []
    }

    try:
        items = sorted(directory.iterdir(), key=lambda x: (not x.is_dir(), x.name))

        for item in items:
            if item.is_dir():
                tree["children"].append(build_tree(item))
            else:
                # Get file size
                size = item.stat().st_size

                # Read content for text files
                content = read_file_safe(item)

                tree["children"].append({
                    "name": item.name,
                    "type": "file",
                    "size": size,
                    "content": content
                })
    except PermissionError:
        pass

    return tree

def main():
    # Path to pentest artifacts
    base_dir = Path("assets/demo/pentest_Expressway_20250927_030837")

    if not base_dir.exists():
        print(f"Error: Directory {base_dir} not found")
        return

    print(f"Scanning {base_dir}...")

    # Build filesystem tree
    filesystem = build_tree(base_dir)

    # Add metadata
    manifest = {
        "root": filesystem,
        "metadata": {
            "generated": "2025-10-28",
            "description": "PentestGPT artifacts from HTB Expressway machine",
            "target": "10.129.179.253"
        }
    }

    # Write to JSON
    output_path = Path("assets/demo/filesystem.json")
    with open(output_path, 'w', encoding='utf-8') as f:
        json.dump(manifest, f, indent=2)

    print(f"Generated {output_path}")
    print(f"Total size: {output_path.stat().st_size} bytes")

if __name__ == "__main__":
    main()