config.json

{
    "max_size": 104857600,
    "max_time": 3600,
    "sleep": 60,
    "database_file": "data/scrape.db",
    "github_token": "github_api_token",
    "local_path": "",
    "save": true,
    "cert_file": "cert.pem",
    "key_file": "key.pem",
    "regexes": [
        {
            "regex": "^[a-zA-Z0-9+_.-]+@[a-zA-Z0-9.-]+\\.[a-zA-Z]{2,}:[^~\/$|: \n]+",
            "bucket": "creds",
            "match": "all"
        },
        {
            "regex": "[a-zA-Z0-9+_.-]+@[a-zA-Z0-9.-]+\\.[a-zA-Z]{2,}",
            "bucket": "email",
            "match": "all"
        },
        {
            "regex": "(?s)-----BEGIN (RSA|DSA|) PRIVATE KEY.*END (RSA|DSA|) PRIVATE KEY-----",
            "bucket": "privkey",
            "match": "all"
        },
        {
            "regex": "\\$[0-9]\\$[a-zA-Z0-9]+\\$[a-zA-Z0-9./=]+",
            "bucket": "pwhash",
            "match": "all"
        },
        {
            "regex": "[a-zA-Z0-9]+::[a-zA-Z0-9]{10}:[a-z0-9]{32}:[a-z0-9-]+",
            "bucket": "pwhash",
            "match": "all"
        },
        {
            "regex": "[a-zA-Z0-9-_]+:[0-9]+:[a-z0-9]{32}:[a-z0-9]{32}",
            "bucket": "pwhash",
            "match": "all"
        },
        {
            "regex": "CVE-[0-9]{4}-[0-9]{4,5}",
            "bucket": "exploit",
            "match": "one"
        }
    ],
    "keywords": [
        {"keyword": "`password`", "bucket": "sqlpass"},
        {"keyword": "proof of concept", "bucket": "exploit"},
        {"keyword": "remote code execution", "bucket": "exploit"},
        {"keyword": "fullz", "bucket": "carder"}
    ]
}