Gias Uddin
Titles as the sources gave them. In 2 entries, with 0 checked quotations.
In the Base
Every entry they appear in.
Then there is prompt injection, a malicious email, ...: The Wild Wide West of OpenClaw Agentic System Security
According to the abstract, Akhond, Khatun and Uddin analyze Reddit posts and security-tagged GitHub issues about the OpenClaw assistant framework. They report 521 weakness instances in 59 types, mostly access control, data exposure, injection and unsafe execution, plus seven recurring mitigation strategies.
How Effectively Do Security Skills Guide Coding Agents in Vulnerability Detection in Linux Kernel?
According to the abstract, Haoran Xue, Nikta Akbarpour, Gias Uddin and Song Wang studied 37 security skills used with Claude Code and four underlying models on 100 real Linux kernel vulnerabilities, 14,800 runs in all. They report 24.4 percent of runs located the correct function, and propose a self-improving skill framework.