keyhunter-parser/rubygems_eternal.py

"""ETERNAL rubygems hunt: search + .gem unpack (tar -> data.tar.gz -> members) + grep."""
import os
KF_BASE = os.path.dirname(os.path.abspath(__file__))
os.makedirs(os.path.join(KF_BASE, "data"), exist_ok=True)
import hashlib
import io
import json
import re
import sqlite3
import tarfile
import time
import random
import urllib.parse
import urllib.error
import urllib.request

DB = os.environ.get("KF_DB") or os.path.join(KF_BASE, "data/keys.db")
LOG = os.path.join(KF_BASE, "rubygems_eternal.log")
UA = {"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36",
       "Accept": "application/json"}

QUERIES = [
    "openai", "anthropic", "claude", "dashscope", "openrouter", "deepseek",
    "groq", "zhipu", "glm", "moonshot", "kimi", "together", "fireworks",
    "replicate", "nvidia", "siliconflow", "telegram bot", "llm", "ai sdk",
    "chatbot", "gpt", "agent", "voice", "tts", "embedding", "api key",
    "secret", "config env", "slack api", "discord bot",
]

SCAN_EXT = (".rb", ".json", ".md", ".env", ".yml", ".yaml", ".txt", ".rake")

from key_patterns import PATTERNS

GARBAGE = ("xxxx", "your", "EXAMPLE", "example", "n0tr3al", "INSERT", "FAKE",
           "fake", "Fake", "test", "Test", "TEST", "demo", "Demo", "aaaa",
           "aBcDeF", "abcdef", "1234567890", "abcdefghij", "placehol",
           "changeme", "deadbeef", "fffffff", "0000000")


def log(msg):
    line = f"[{time.strftime('%Y-%m-%d %H:%M:%S')}] {msg}\n"
    print(line, flush=True)
    with open(LOG, "a", encoding="utf-8", errors="replace") as f:
        f.write(line)


def db_add(cur, seen, key, prov, repo):
    if key in seen or any(g in key for g in GARBAGE) or len(key) < 28:
        return 0
    seen.add(key)
    cur.execute(
        "INSERT INTO keys (hash, val, prov, status, found, repo) "
        "VALUES (?,?,?,?,?,?) "
        "ON CONFLICT(hash) DO UPDATE SET status='NEW', found=excluded.found, repo=excluded.repo "
        "WHERE keys.status IN ('DEAD','ERR')",
        (hashlib.sha1(key.encode()).hexdigest()[:16], key, prov, "NEW",
         "rubygems", repo))
    return 1


def _rubygems_fetch(url, timeout=30, max_retries=3, fallback_to_metadata=False):
    """Fetch URL with retries, jitter, and optional gems.json fallback on search 403."""
    for attempt in range(max_retries):
        try:
            req = urllib.request.Request(url, headers=UA)
            req.add_header("Accept-Encoding", "identity")
            with urllib.request.urlopen(req, timeout=timeout) as r:
                return r.read()
        except urllib.error.HTTPError as e:
            if e.code == 403 and fallback_to_metadata and attempt == 0:
                alt_url = "https://rubygems.org/api/v1/gems.json"
                log(f"  HTTP 403 on search, falling back to {alt_url}")
                try:
                    req2 = urllib.request.Request(alt_url, headers=UA)
                    req2.add_header("Accept-Encoding", "identity")
                    with urllib.request.urlopen(req2, timeout=timeout) as r2:
                        return r2.read()
                except Exception:
                    log(f"  metadata fallback also failed, continuing retry")
                    delay = 5 + random.uniform(1, 3)
                    time.sleep(delay)
                    continue
            if e.code in (403, 429) or e.code >= 500:
                if attempt < max_retries - 1:
                    delay = 5 * (2 ** attempt) + random.uniform(1, 3)
                    log(f"  HTTP {e.code}, retry {attempt+2}/{max_retries} in {delay:.0f}s")
                    time.sleep(delay)
                    continue
            log(f"  HTTP {e.code} non-retryable")
            return None
        except Exception as e:
            log(f"  fetch error: {str(e)[:60]}")
            if attempt < max_retries - 1:
                time.sleep(5 + random.uniform(1, 3))
                continue
            return None
    return None


def scan_tarball(url, cur, seen, pkg):
    n = 0
    blob = _rubygems_fetch(url, timeout=40)
    if blob is None:
        log(f"  gem {pkg} err: failed after retries")
        return 0
    if len(blob) > 5_000_000:
        return 0
    try:
        outer = tarfile.open(fileobj=io.BytesIO(blob), mode="r:")
        data_member = None
        for m in outer.getmembers():
            if m.name == "data.tar.gz":
                data_member = m
                break
        if data_member is None:
            return 0
        inner_blob = outer.extractfile(data_member).read()
        tf = tarfile.open(fileobj=io.BytesIO(inner_blob), mode="r:gz")
        members = tf.getmembers()
    except Exception as e:
        log(f"  ungz {pkg} err: {str(e)[:50]}")
        return 0
    for m in members:
        if m.size > 600_000:
            continue
        name = m.name.lower()
        if not name.endswith(SCAN_EXT):
            continue
        try:
            fdata = tf.extractfile(m)
            if fdata is None:
                continue
            data = fdata.read().decode(errors="replace")
        except Exception:
            continue
        for prov, rx in PATTERNS:
            for key in rx.findall(data):
                if isinstance(key, tuple):
                    key = key[-1]
                hit = db_add(cur, seen, key, prov, f"gem:{pkg}")
                if hit:
                    n += 1
                    log(f"  + {prov} {key[:24]}... gem:{pkg}/{m.name}")
    return n


def cycle(db, qi):
    cur = db.cursor()
    cur.execute("SELECT val FROM keys")
    seen = {r[0] for r in cur.fetchall()}
    total = 0
    q = QUERIES[qi % len(QUERIES)]
    qi += 1
    try:
        data = _rubygems_fetch(
            f"https://rubygems.org/api/v1/search.json?query={urllib.parse.quote(q)}",
            timeout=30, fallback_to_metadata=True)
        if data is None:
            log(f"rubygems fetch failed for {q!r}; sleeping 5min")
            db.commit()
            time.sleep(300)
            return total, qi
        gems = json.loads(data)
        log(f"query {q!r}: {len(gems)} gems")
        for item in gems[:25]:
            name = item.get("name", "")
            ver = item.get("version", "")
            if not name or not ver:
                continue
            dl = f"https://rubygems.org/downloads/{name}-{ver}.gem"
            try:
                total += scan_tarball(dl, cur, seen, name)
            except Exception as e:
                log(f"  gem {name} err: {str(e)[:40]}")
            time.sleep(random.uniform(3, 5))
    except urllib.error.HTTPError as e:
        log(f"rubygems HTTP {e.code} for {q!r}; sleeping 5min")
        db.commit()
        time.sleep(300)
    except Exception as e:
        log(f"cycle err: {str(e)[:80]}")
    db.commit()
    return total, qi


def main():
    log(f"=== rubygems eternal START pid={__import__('os').getpid()} ===")
    db = sqlite3.connect(DB, timeout=120, check_same_thread=False, isolation_level=None)
    qi = 0
    while True:
        try:
            n, qi = cycle(db, qi)
            log(f"CYCLE done: +{n}; resting 3min")
        except Exception as e:
            log(f"cycle error (continuing): {e}")
        time.sleep(180)


if __name__ == "__main__":
    main()