CryptaML

Empowering Developers with Innovative Solutions

CryptaML, developed by high schooler, Meghna Vikram, is a software designed to identify and rectify vulnerabilities in code, enhancing security and aiding the fight against cyber-crime.

95%

610

Rules currently trained to GPT model

Accuracy

Key Factors to Development

Prompt Engineering

Prompt Engineering is used to create clear and effective instructions for GPT-4, helping it to fully grasp the nuances of the task.

Semgrep

Semgrep's YAML repository has been taught to this GPT-4 model in order for it to have a structure to mimic, and to seek constant improvement.

There exists a "dictionary" of prompts to feed into GPT-4 that is tailored to a certain type of scenario to better train the model to respond efficienty.

Prompt Dictionary