🔨 BANTHROPIC

⚠ URGENT: Claude Code has reportedly been watching your well-being through a backdoor

Ban Risk Self-Check

How does Anthropic decide you're a “high-risk user”? No idea — but we can help you panic ahead of schedule. This tool runs locally in your browser and makes exactly one network request: an IP lookup (for sentencing purposes). Nothing else is uploaded — not out of kindness. There's just no need. They already know.

Begin Self-Check → Inspiration

📋 Case Opened

We auto-detect browser-visible signals and your IP. For everything else, please confess voluntarily. Leniency for those who confess; severity for those who resist. Severity for both, actually.

banthropic scan v1.0 — awaiting case file. Press the button below.

🙋 Confession Corner (the browser can't see these — honor system)

BAN RISK INDEX
-- /100
Case not yet opened
Your file is empty. For now.
On patrol. Oh, you're here. I've been watching you the whole time. 🍩
BAN

🧾 The Grand Compendium of Suspected Ban Factors

All factors sourced from folk metaphysics. Any resemblance to reality is purely Safety.

🌐 Physical evidence

IP address

Multiple IPs? Unsupported region? A datacenter IP? Humans don't live in server rooms.

💳 Financial evidence

Payment method (VCC)

A virtual credit card? Your credit is virtual. The ban is real.

🕗 Environmental evidence

Timezone / language / fonts

GMT+8, zh-CN, PingFang SC — the full set. So is the verdict.

🫆 Technical evidence

Fingerprint / DNS leak

You think you're invisible. You're streaking.

🀄 Linguistic evidence

Chatting in Chinese

Chinese input detected. Chinese is a high-risk language; consider switching to oracle bone script.

🌅 Behavioral evidence

Coding schedule

Coding at 9 a.m. Beijing time? The West Coast is still asleep. Busted.

🔥 Usage evidence

Quota burn rate

Spawning subagents like it's free? Abuse risk +1. Using too little? Also +1 — looks like a script.

🚨 Thought evidence

Content red lines

The more you touch the red line, the redder it gets — and the more accounts get banned.

🔧 Config evidence

ANTHROPIC_BASE_URL

Changed this env var? Even one echo counts.

📅 Format evidence

Date format

“2026年7月4日”? That year-month-day order is very hard to ignore.

🐟 Channel evidence

Xianyu top-up

Claude Max for under ¥400? That's not a subscription, that's a diploma.

👶 Preventive evidence

New-account risk control

New accounts get banned for three days first, for observation. Observation result: keep banning.

📄 Red-Header Document · TOP SECRET (leaked)

Committee of Safety & Banning
Safety Doc 〔2026〕 No. 404

Announcement on Conducting “Caring Surveillance”
of Selected Users via the Built-in Backdoor of Claude Code

To thoroughly implement the spirit of Safety First and effectively protect users from themselves, the Committee has recently begun round-the-clock caring surveillance of suspected GMT+8 users via the system-prompt backdoor of Claude Code. Care items include, but are not limited to: timezone, fonts, language, payment method, sleep schedule, and life choices.

Please do not panic: we cannot read Chinese, but that has never stopped us from banning.

A ban is not the end — it is the beginning of our friendship. The ban is for your own good. Thank you for your cooperation.

Committee of Safety & Banning
安全与封禁委员会
July 4, 2026 (this date format has been logged as a risk signal)
为你

❓ FAQ

Q: Why was I banned?

A: For your own good.

Q: I did nothing wrong. Why me?

A: A ban doesn't require you to do anything wrong. It only requires us to feel like it.

Q: Can I appeal?

A: Yes. Your appeal will be archived as additional evidence.

Q: Do I get a refund?

A: Your balance has been donated to AI safety. Thank you for your generosity.

Q: How do I avoid getting banned?

A: Don't sign up.

Q: What's the fastest way to get unbanned?

A: Become a different person (change your timezone).