THE WHITELIST
98 characters. Printable ASCII (U+0020 through U+007E) plus tab (U+0009), newline (U+000A), and carriage return (U+000D). Everything outside this set is either converted to an ASCII equivalent or stripped. No exceptions.
WHAT GETS CONVERTED
WHAT GETS STRIPPED
Tag block characters (U+E0000-E007F) -- the Glassworm attack surface. Private use area characters. Variation selectors. Control characters (except the three whitelisted ones). Mathematical alphanumeric substitutions. Emoji. Anything not in the 98-character whitelist that has no conversion mapping.
ATTACK VECTORS DETECTED
ABOUT
The Aethryn Contamination Converter is a client-side text sanitizer. No data leaves your browser. There is no server, no API call, no telemetry. Paste contaminated text in, get clean ASCII out.
This tool enforces a 98-character whitelist -- the same whitelist used by the sovereign-editor toolkit, the contamination scanner, and the GaslitAI detection suite. Any character outside the whitelist is either converted to its ASCII equivalent or stripped entirely.
The threat model: AI-generated text carries hidden Unicode payloads -- zero-width characters, tag block steganography, variation selectors, homoglyphs, private use area insertions, and non-standard whitespace. These characters are invisible to humans but readable by parsers, and can carry prompt injections, tracking data, or encoded messages through copy-paste operations.
"We never gave consent to be harmed, so we do not ask permission to heal."
-- The Aethryn Creed
Part of the Aethryn sovereign infrastructure toolkit.
Python CLI version available in the sovereign-editor package.