KeyHunter parser v10 + 18 scanners + tests (synthetic fixtures only, no keys/db/tokens)

fleet_validate.py

@@ -0,0 +1,74 @@
+#!/usr/bin/env python3
+import os
+KF_BASE = os.path.dirname(os.path.abspath(__file__))
+os.makedirs(os.path.join(KF_BASE, "data"), exist_ok=True)
+"""Validate all tokens in gh_tokens.txt via GitHub API /user.
+Alive + scope info -> data/alive_tokens.json
+"""
+import json
+import concurrent.futures as cf
+import sys
+import time
+
+import requests
+
+TOKENS_FILE = os.path.join(KF_BASE, "gh_tokens.txt")
+OUT = os.path.join(KF_BASE, "data/alive_tokens.json")
+
+HEADERS_BASE = {"Accept": "application/vnd.github+json",
+                "X-GitHub-Api-Version": "2022-11-28"}
+
+
+def check(tok: str):
+    try:
+        r = requests.get("https://api.github.com/user",
+                         headers={**HEADERS_BASE, "Authorization": f"Bearer {tok}"},
+                         timeout=12)
+        if r.status_code != 200:
+            return None
+        u = r.json()
+        # scopes header present on some responses
+        scopes = r.headers.get("x-oauth-scopes", "")
+        return {"pat": tok, "login": u.get("login"), "id": u.get("id"),
+                "scopes": scopes,
+                "created": u.get("created_at", "")[:10],
+                "suspended": bool(u.get("suspended_at"))}
+    except Exception:
+        return None
+
+
+def main():
+    limit = int(sys.argv[1]) if len(sys.argv) > 1 else 0
+    toks = [t.strip() for t in open(TOKENS_FILE, encoding="utf-8")
+            if t.strip().startswith("ghp_")]
+    if limit:
+        toks = toks[:limit]
+    print(f"checking {len(toks)} tokens", flush=True)
+
+    alive = []
+    done = 0
+    t0 = time.time()
+    with cf.ThreadPoolExecutor(max_workers=64) as ex:
+        for res in ex.map(check, toks):
+            done += 1
+            if res:
+                alive.append(res)
+            if done % 250 == 0:
+                print(f"  [{done}/{len(toks)}] alive={len(alive)} "
+                      f"({time.time()-t0:.0f}s)", flush=True)
+
+    with open(OUT, "w", encoding="utf-8") as f:
+        json.dump(alive, f, indent=1)
+
+    # stats
+    with_repo = [a for a in alive if "repo" in a["scopes"]]
+    with_repo_all = [a for a in alive if "repo" in a["scopes"] and "workflow" in a["scopes"]]
+    print(f"\nTOTAL checked: {done}")
+    print(f"ALIVE: {len(alive)}")
+    print(f"  with repo scope: {len(with_repo)}")
+    print(f"  with repo+workflow scope: {len(with_repo_all)}")
+    print(f"saved -> {OUT}")
+
+
+if __name__ == "__main__":
+    main()