Cutting-
Edge,
Security
AI.

Protosec

[[Prototype]]

is a Threat-targeting Researching Lab focusing on the world's most cutting-edge fields.

Artificial Intelligence

Machine Learning

Security

AI/ML

Natural-Language-Processing

Machine-Learning based

Exploitation Automation

AutoPwn

Threat Identification

Protosec has made significant impacts on the world by identifying real-life threats.

Transformers Model-Deserialization Remote-Code Execution

The load_repo_checkpoint() function under the TFPreTrainedModel() class. This vulnerability enables attackers to execute arbitrary code and commands by using a carefully crafted serialized payload.

Llama.cpp RPC Heap-overflow Remote-Code Execution

A carefully crafted overflow exploiting Tensor dimension calculations enables arbitrary control over internal memory structures. By partial pointer overwrites, meticulous heap structuring, and Structure-Oriented Programming, bypassed multiple layers of sanitization checks

Llama-Cpp-Python Remote Code Execution by Server-Side Template Injection in Model Metadata

Remote-Code Execution due to Server-Side Template Injection of unportection renderer behaviour in GGUF Model Metadata in Llama-cpp-Python

Future of Security

Discover how AI is revolutionizing
the security landscape.

Active Threat Intelligence

Active Research on leading AI/ML Projects

ML Security Automation

ML-based security automation for threat-intelligence, reverse engineering, dynamic debugging, and vulnerability detection.

LLM Offensive Security

Offensive Security targeting fragility of AI/ML models, diving deep into the inner workings of LLMs

Our Projects

Pioneering AI-driven security tools
for next-generation challenges

Join Our Team

Shape the future of
AI/ML security research