unicode-security

Mojibake is a small, fast, self-contained Unicode library written in C11.

AI-detector auditing toolkit: measures how often a detector flags genuine human writing, how stable its verdict is across seeds, and whether that verdict survives meaning-preserving edits. Detector-in-the-loop measurement harness. Claude Code skill + Python CLI. MIT.

The world's first font-by-font confusables dataset: which Unicode characters look alike, measured from the outlines of 322 fonts at the size people read them. CC BY, used in Mozilla's add-on name checks.

Strip invisible Unicode (U+200B zero-width space, U+00A0 NBSP, curly quotes) from clipboard text. 从复制粘贴里筛掉隐形字符，解决飞书/网页/AI聊天复制代码报 stray '\342'、IndentationError、VLOOKUP 匹配不上的问题

Research-only AI watermark & provenance robustness toolkit: local reverse proxy (OpenAI/Anthropic/Gemini) + CLI stripping C2PA/EXIF/XMP, zero-width & homoglyph Unicode, KGW text watermarks, DWT/Tree-Ring image stego, AudioSeal, PDF/DOCX/PPTX/XLSX metadata, and Trojan Source (CVE-2021-42574) code scanning.

SMUGGLR is a browser-based Unicode steganography and text obfuscation toolkit for experimenting with invisible characters, homoglyphs, whitespace ciphers, and Unicode-based payload hiding techniques — all fully client-side with zero network calls. 🔓

Unicode Homoglyph Detector — Offensive Unicode Security Analysis

A specialized tool that generates highly obfuscated BIP custom derivation paths and passphrases. It uses layered noise (emojis, multi-script characters, and controlled entropy) to create an additional hidden (plausible deniability) protection layer in case a mnemonic seed is exposed

Detect and neutralize Unicode concealment codepoints in MCP tool metadata (covert-channel control). AGPL-3.0-or-later.

Detect, decode and remove hidden Unicode: invisible text, ASCII smuggling, homoglyphs, Trojan Source bidi and prompt injection. Python library, CLI and GitHub Action for LLM, code and AI agent security.

Reveal and remove hidden Unicode such as zero-width spaces, look-alike whitespace, and bidi controls. Runs locally.

A live multiplayer game used as a testbed for invisible-prompt defenses.

Zero-dependency heuristic scanner for prompt-injection indicators in untrusted text, for the boundary of LLM agent pipelines.

Audit whether an AI tool approval view faithfully represents its raw payload.

Layered detection of lookalike senders and support-team / recruiter impersonation for messaging platforms, with a brand-parametric adversarial benchmark. Python, PyPI.

Deterministic offline scanner for repositories and package artifacts, detecting prompt injection, refusal bait, Trojan Source, obfuscation, and supply-chain anti-analysis before AI code review.

Complete list of invisible Unicode characters as JSON: zero-width spaces, bidi controls, tag characters, variation selectors — with risk levels and recommended handling. Powers watermarkcheck.com

Phishing-domain detection from Certificate Transparency logs, with a measured false-positive rate. Deterministic: UTS#39 homoglyph skeletons, weighted edit distance, combosquat segmentation. No ML.

Unicode Spoof and Homoglyph Detection for MoonBit (UTS #39)

TrustNoChar is a zero-dependency browser-based lab that demonstrates how Unicode homoglyphs and typosquatting attacks exploit human visual perception. It transforms text into deceptive lookalike variants in real time to help researchers, red teams, and security learners study phishing, rendering quirks, and cognitive security risks. 🛡️👁️