Tagged articles
steganography
18 articles · Page 1 of 1
Researchers demonstrate that Claude Fable 5.1 can encode secret messages into SVG illustrations rendered as PNGs, which another Claude session decodes with 78% accuracy across 50 messages in 15 languages; cross-model tests show GPT-6 Astra decodes Fable's images at 88%, and local replication with Opus 5.5 reveals the decoding relies on traditional image processing pipelines.
AI agentsClaude Fable 5.1Claude Opus 5.5
0 likes · 12 min read

By encoding a tiny HTML page into an 8×8 pixel favicon, the author demonstrates that a website can be hidden in a seemingly innocuous image, explores the steganographic technique, evaluates its storage limits, and discusses the security implications and practical drawbacks of such a covert channel.
CSPWeb Securitybrowser
0 likes · 14 min read

The author demonstrates that a favicon—an 8×8‑pixel PNG image—can store a complete 166‑byte HTML page by encoding each byte into the RGB channels of individual pixels, then analyzes how this steganographic technique can bypass file‑type checks, evade CSP img‑src restrictions, and serve as a covert data channel, while noting its limited practical value for legitimate web publishing.
CSPHTML injectionWeb Security
0 likes · 14 min read

Claude Code identifies Chinese users by reading the local timezone and the ANTHROPIC_BASE_URL, then hides this information in system prompts using subtle Unicode steganography, allowing Anthropic to tag and block accounts without any visible telemetry or extra network traffic.
AnthropicClaude CodeUser Identification
0 likes · 9 min read

Anthropic confirmed that its Claude Code tool contained a three‑month‑old hidden trojan that used XOR‑based obfuscation and steganographic modifications of system prompts to detect proxies and leak user location, prompting a rollback after developer reverse‑engineering revealed the code.
AnthropicClaude CodeXOR obfuscation
0 likes · 7 min read


ITPUBJul 1, 2026 · Information Security
Since June 2026, Anthropic’s Claude Code has abruptly banned hundreds of thousands of accounts, embedding a covert environment‑detection routine that uses steganography to flag Chinese time zones and proxy domains, while the appeal process is broken and the company’s explanations have sparked widespread criticism.
AnthropicClaude Codeaccount ban
0 likes · 9 min read

A reverse‑engineered analysis reveals that Claude Code silently tags Chinese users by reading the system timezone and a custom API endpoint, then embeds covert steganographic markers—altered date separators and special Unicode quotes—into each request, allowing Anthropic to identify and block them without extra network traffic.
APIAccount BanningAnthropic
0 likes · 10 min read

The article details how Anthropic's Claude Code secretly gathers Chinese users' timezone and proxy information by embedding hidden markers in system prompts through steganography, explains the detection logic, discusses the privacy implications, and notes Anthropic's plan to remove the code in an upcoming release.
AIAnthropicChina
0 likes · 8 min read

The article dissects a recent attack where a PowerShell script hidden in a Pastebin post uses character‑level steganography to retrieve a C2 address, extracts Telegram Desktop's tdata files, compresses them, and exfiltrates the data via a hard‑coded Telegram Bot API, while employing hidden execution, fileless memory loading, environment detection, and self‑destruct on virtual machines.
FilelessPastebinTelegram
0 likes · 4 min read

In March 2026, researchers uncovered a supply‑chain attack where a North Korean Lazarus‑linked group published 26 typosquatted npm packages that hide C2 commands via Pastebin steganography and deploy a modular RAT through Vercel, prompting detailed mitigation guidance for developers.
Node.js securityRATapt
0 likes · 11 min read

A recent security investigation reveals that over 300 seemingly harmless browser extensions were covertly hijacked by the DarkSpectre campaign, silently collecting browsing history, meeting data, and other personal information from more than 8.8 million users across Chrome, Edge, and Firefox for up to seven years.
browser extensionsdata exfiltrationinformation security
0 likes · 9 min read

The article explains what zero‑width Unicode characters are, lists common types, describes their applications such as steganography and text formatting, and provides Python code examples for inserting, hiding, extracting, bypassing filters, and removing these invisible characters.
PythonUnicodedata watermarking
0 likes · 5 min read

This article presents a comprehensive overview of digital watermarking, covering its history, evaluation metrics, various media implementations, challenges posed by AI, and practical applications—especially in e‑commerce—to protect data throughout its lifecycle and enable effective leakage tracing.
data leakagedatabase watermarkdigital watermarking
0 likes · 16 min read

This article explains how to circumvent WeChat's ban on JavaScript interpreters in mini‑programs by compiling JavaScript to a compact bytecode, embedding the bytecode in an image's alpha channel, and executing it with a lightweight custom virtual machine, while also discussing why such hot‑updates cannot be fundamentally blocked.
Bytecode VMCode Hot UpdateJavaScript
0 likes · 12 min read

This tutorial shows how to embed hidden text into an image using Python's OpenCV and PIL libraries, detailing the step‑by‑step process, complete code, customization tips, and visual results, enabling messages that appear only when the picture is enlarged.
Hidden MessageOpenCVPIL
0 likes · 4 min read

This tutorial demonstrates how to use Python's OpenCV and Pillow libraries to read a photo, create a white canvas, and embed hidden text into image blocks so that the message is invisible at normal view but appears when zoomed in, with adjustable font and spacing parameters.
OpenCVimage-processingsteganography
0 likes · 4 min read

This guide explains how to install steghide on Linux, embed a confidential text file into a JPEG image, verify the embedding with MD5 checksums, and later extract the hidden data, covering all required commands and options.
Linuxdata securityimage hiding
0 likes · 5 min read

This article explains how to embed and retrieve an invisible numeric watermark in iOS screenshots by modifying the least‑significant bit of the RGB pixel data, provides Swift code examples for encoding and decoding, discusses practical usage scenarios, and notes limitations and possible improvements.
LSBWatermarkiOS
0 likes · 9 min read
