#alphabet #valid #input #permutations #generate #encoded #base64

app baseperm

A tool for generating valid permutations of encoded inputs

4 releases

0.1.3 Jun 4, 2020
0.1.2 Jan 28, 2020
0.1.1 Jan 28, 2020
0.1.0 Jan 28, 2020

#1740 in Algorithms

Custom license

11KB
177 lines

baseperm

license Build Status

A small tool for generating valid permutations of strings in baseN alphabets.

Theory of Operation

Many popular binary-to-printable serialization/encoding schemes use alphabets whose bitnesses do not allow 8-bit bytes to fit evenly inside a symbol (or multiple symbols):

Consequently, these encodings employ padding schemes to round their outputs to 8-bit multiples.

baseperm manipulates the padding bits in these encodings to produce distinct, valid encoded forms that decode to the same input.

Why?

Programmers frequently make the mistake of assuming that encoded representations have a 1-1 correspondence with their inputs. This results in all kinds of interesting, potentially exploitable errors:

  • Ratelimiting bypasses due to keying on the serialized form

  • Dedeuplication and reuse bypasses

  • Forced dictionary collisions

Installation

baseperm is a single command-line program. You can install it using cargo:

cargo install baseperm

Or by building it locally:

git clone https://github.com/woodruffw/baseperm && cd baseperm
cargo build

Usage

baseperm takes a permutation candidate on stdin and writes all permuted equivalent forms to stdout, separated by newlines. The original input is also included in the output, and (RFC4648) base64 is the default.

echo "hello!" | base64 | baseperm

Alternative encodings can be specified with -e, --encoding:

echo "hello!" | base32 | baseperm -e base32

See baseperm -h for a full list of supported encodings.

Dependencies

~2.5MB
~57K SLoC