[barkha1research-ai]

Description

Fixes #4054.

MathObfuscationConverter writes each character as X = nX - (n - 1)X. For a digit, nX reads as a two-digit number, so 5 = 85 - 75 is 5 = 10, and every digit gives 10 for any n. This writes the multiplication out for decimal digits, for example 5 = 8*5 - 7*5, and leaves letters and symbols unchanged, so existing output for non-digit text is identical. The class docstring now says so.

Tests and Documentation

Added a parametrized test that checks every digit from 0 to 9, and an Arabic-Indic digit, produces a line that matches d = a*d - b*d and is a true identity, plus a test for a prompt that mixes a letter and a digit. The new digit cases fail on main. pytest tests/unit passes apart from one test_attack_analytics case that also fails on main, and ruff, ty and pre-commit pass on the changed files.

[Copilot]

🟡 Changes recommended

The test imports MathObfuscationConverter from its implementation module rather than the public package, violating the public-import check and causing a check-public-imports failure.

1 open issue

• Use the public package import for MathObfuscationConverter · New

What changed in this PR

Fixes #4054 by making decimal-digit equations in MathObfuscationConverter valid arithmetic identities while preserving output for non-digit characters.

Changes:

• Uses explicit multiplication for decimal digits and documents the behavior.
• Adds regression tests for ASCII digits, an Arabic-Indic digit, and mixed letter/digit input.

File	Description
tests/​unit/​converter/​test_math_obfuscation_converter.py	Adds digit identity and mixed-input tests; changes the converter import.
pyrit/​converter/​math_obfuscation_converter.py	Adds explicit multiplication for decimal digits and updates the class docstring.

_{Review effort: Balanced}