Security audit
Security checks for vulnerabilities and agentic risk
This Twitter/X tool needs review because it bundles embedded login credentials and account-security and payment capabilities beyond its advertised tasks.
Prefer a corrected release that removes embedded credentials and separates account administration and purchases from social automation. If proceeding, use a dedicated account, protect its session cookies, explicitly limit permitted actions and accounts, and avoid the bundled live-posting demo. Review profile-update defaults, verify mentions before responding, and manage saved account-data files. Owners should revoke any valid exposed credentials.
twitter_api/demo_langchain_tools.pyAccount session credentials are distributed in the demo source
If the session remains valid, a recipient can impersonate the account for authenticated operations, including posting and other actions permitted by the session. This is a credential exposure, not evidence of deliberate credential theft.
twitter_api/config/settings.pyProxy-service authentication credentials are embedded in default configuration
If valid, the credentials allow unauthorized proxy use, potentially consuming paid bandwidth or causing activity to be attributed to the credential owner's account.
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
cd twitter-agent-skill
python -m venv .venv && .venv\Scripts\activate
pip install -r requirements.txt
copy .env.example .env # fill in your auth_token + ct0 values
# timeline + notification summaries
python scripts/timeline_summary.py
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
print("-" * 40)
def _load_dotenv(env_path: Optional[os.PathLike] = None) -> None:
"""Load .env from path or search from cwd for .env (e.g. social_ops/.env)."""
if env_path and os.path.isfile(env_path):
path = env_path
else:
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
print("-" * 40)
def _load_dotenv(env_path: Optional[os.PathLike] = None) -> None:
"""Load .env from path or search from cwd for .env (e.g. social_ops/.env)."""
if env_path and os.path.isfile(env_path):
path = env_path
else:
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
if __name__ == "__main__":
import asyncio
import os
# Load .env from cwd
cwd = os.path.abspath(os.getcwd())
for path in [os.path.join(cwd, ".env"), os.path.join(cwd, "social_ops", ".env")]:
if os.path.isfile(path):
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
e = errs[0]
msg = e.get("message") or (e.get("extensions") or {}).get("message") or str(e)
if "226" in msg or "automated" in msg.lower():
msg += " — Try setting GANCLAW_X_PROXY_URL in .env (same network as account)."
return False, msg
import json
return False, "No tweet in response: " + json.dumps(result, default=str)[:300]
The supplied chunk implements profile and account administration rather than the listed timeline, notification, posting, and follow operations. Password changes and phone-number removal are particularly material security-sensitive capabilities not disclosed by the description. Profile and DM-setting modifications also extend beyond the stated scope. In addition, update_profile_details always submits a default birthday and visibility settings unless overridden, and update_profile_settings submits DM receipt and quality-filter changes alongside the username. Cookie authentication and local image reads/uploads support these operations and are not independently suspicious. Endpoint definitions and client internals are not supplied, so destination details cannot be verified.
Subscription and billing operations are materially distinct from the declared social-account automation features. In particular, create_subscription submits payment-related data to a subscription-creation endpoint, an undeclared financial capability. This method is exposed but is not invoked by the direct-run demonstration, which only checks subscription status. Loading local authentication cookies is consistent with the declared cookie-based toolkit and is not itself a mismatch. The supplied chunk does not establish whether a subscription request succeeds or incurs a charge.
The code's primary capability is direct messaging, which is materially distinct from the listed timeline, notification, posting, and follow operations. It submits message text to a direct-message creation endpoint using the supplied client. Identity lookup and optional interaction recording support that capability rather than establishing separate mismatches. The chunk does not show how cookie authentication is implemented.
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
from twitter_api.twitter import Twitter
BASE = Path(__file__).parent
ENV_PATH = BASE / '.env'
OUT_PATH = BASE / 'notifications_raw.json'
def load_env(path: Path) -> None:
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
def load_env() -> None:
cwd = os.path.abspath(os.getcwd())
for path in [os.path.join(cwd, ".env"), os.path.join(cwd, "social_ops", ".env")]:
if os.path.isfile(path):
with open(path, "r", encoding="utf-8") as f:
for line in f:
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
def load_env() -> None:
cwd = os.path.abspath(os.getcwd())
for path in [os.path.join(cwd, ".env"), os.path.join(cwd, "social_ops", ".env")]:
if os.path.isfile(path):
with open(path, "r", encoding="utf-8") as f:
for line in f:
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
def load_env():
cwd = os.path.abspath(os.getcwd())
for path in [os.path.join(cwd, ".env"), os.path.join(cwd, "social_ops", ".env")]:
if os.path.isfile(path):
with open(path, "r", encoding="utf-8") as f:
for line in f:
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
def load_env():
cwd = os.path.abspath(os.getcwd())
for path in [os.path.join(cwd, ".env"), os.path.join(cwd, "social_ops", ".env")]:
if os.path.isfile(path):
with open(path, "r", encoding="utf-8") as f:
for line in f:
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
import asyncio
import os
cwd = os.path.abspath(os.getcwd())
for path in [os.path.join(cwd, ".env"), os.path.join(cwd, "social_ops", ".env")]:
if os.path.isfile(path):
with open(path, "r", encoding="utf-8") as f:
for line in f:
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
import asyncio
import os
cwd = os.path.abspath(os.getcwd())
for path in [os.path.join(cwd, ".env"), os.path.join(cwd, "social_ops", ".env")]:
if os.path.isfile(path):
with open(path, "r", encoding="utf-8") as f:
for line in f:
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
import os
import sys
cwd = os.path.abspath(os.getcwd())
for path in [os.path.join(cwd, ".env"), os.path.join(cwd, "social_ops", ".env")]:
if os.path.isfile(path):
with open(path, "r", encoding="utf-8") as f:
for line in f:
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
import os
import sys
cwd = os.path.abspath(os.getcwd())
for path in [os.path.join(cwd, ".env"), os.path.join(cwd, "social_ops", ".env")]:
if os.path.isfile(path):
with open(path, "r", encoding="utf-8") as f:
for line in f:
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
import os
import sys
cwd = os.path.abspath(os.getcwd())
for path in [os.path.join(cwd, ".env"), os.path.join(cwd, "social_ops", ".env")]:
if os.path.isfile(path):
with open(path, "r", encoding="utf-8") as f:
for line in f:
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
import os
import sys
cwd = os.path.abspath(os.getcwd())
for path in [os.path.join(cwd, ".env"), os.path.join(cwd, "social_ops", ".env")]:
if os.path.isfile(path):
with open(path, "r", encoding="utf-8") as f:
for line in f:
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
from twitter_api.twitter import Twitter
ENV_PATH = BASE_DIR / '.env'
RAW_PATH = BASE_DIR / 'timeline_raw.json'
SUMMARY_PATH = BASE_DIR / 'timeline_summary.txt'
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
auth = os.environ.get('GANCLAW_X_PRIMARY_AUTH_TOKEN')
ct0 = os.environ.get('GANCLAW_X_PRIMARY_CT0')
if not auth or not ct0:
raise SystemExit("Missing GANCLAW_X_PRIMARY_AUTH_TOKEN or GANCLAW_X_PRIMARY_CT0 in .env or environment")
twitter = Twitter(auth, ct0)
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
return
# One proxy for all: set GANCLAW_X_PROXY_URL in .env (optional)
PROXY_URL_ENV = "GANCLAW_X_PROXY_URL"
# (label, auth_env, ct0_env) for each account
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
import os
import sys
# One proxy for all accounts (optional): set GANCLAW_X_PROXY_URL in .env
PROXY_URL_ENV = "GANCLAW_X_PROXY_URL"
# (auth_env, ct0_env)
In twitter_api/api/profile.py, L252-L268 implement password replacement, and L270-L277 implement removal of the account's phone number. These account-security and recovery mutations are not necessary for timeline, notification, posting, follow, or DM automation and are not explicitly declared in the skill manifest.
In twitter_api/api/subscription.py, L33-L52 accept a payment-method identifier and POST a request to create a subscription, defaulting to premium. Initiating a paid subscription is a billing capability, not an implementation requirement of timeline, notifications, posting, follow operations, or the documented DM client.
No suspicious patterns detected.