2026-07-12 / Signal #5
Anthropic's secret "experiment" hidden tracker in Claude Code (steganography markers in prompts to detect Chinese distillation/abuse, despite anti-surveillance stance; code removed after discovery).
“Anthropic hid a secret tracker in Claude to catch model thieves. Their anti-surveillance users were not amused.”
7.7Weirdness
Why It Matters
Ethical" AI lab using stealth prompt steganography to spy on users/labs; darkly funny hypocrisy and real artifact changes builder trust in tools.
Evidence
Malwarebytes/Ars Technica/Reddit/The Register reports (early July discovery, ongoing discussion + removal in recent HN/surveillance threads).
Caveat
Not brand-new today; use only if making a trust-in-dev-tools episode or tying to a fresh Anthropic update.
Signal Read
Receipts: 8Story voltage: 9Heat: 6
Source Trail
Daily scan: 2026-07-12
- No public source URL captured yet.