Systems Security · Fuzzing · Web3 · AI Agents

I'm a master's student in Electronic Information at Tianjin University, with a background in cyberspace security. I work on software security, particularly where operating systems, fuzzing, and program analysis meet.

My current focus is understanding how low-level systems fail: reproducing kernel bugs, tracing them to their root causes, and turning findings into fixes or better testing approaches. I also explore smart-contract security and how LLMs and agents can assist with auditing, testing, and security research.

• Kernel & filesystem security — Linux and Rust-based operating systems, filesystem fuzzing, memory-safety issues, permission semantics, and syscall behavior.
• Web3 security — Solidity contract analysis, vulnerability detection, property-based fuzzing, and LLM-assisted auditing.
• AI & agent engineering — security evaluation of agent tools, automated analysis workflows, and practical LLM infrastructure.

I contribute bug reports and patches to upstream systems projects:

• Asterinas — authored merged fixes for ownership changes, sticky-bit enforcement, and supplementary-group permission checks.
• Linux kernel — reported an NTFS bug credited in a mainline fix, and submitted HFS filesystem patches reviewed and tested upstream.

I care about small, reproducible bug reports, regression tests, and patches that preserve expected system behavior.

• web3scan — an AI-assisted smart-contract vulnerability scanning tool.
• llm4echidna — exploring LLM-generated invariants and Echidna-based smart-contract fuzzing.
• skills-security-benchmark — a curated research corpus for agent/tool-use security evaluation.
• crypto-market-skills — composable agent skills for crypto-market data and research workflows.
• AI API directory — an open directory of AI API services and pricing references.

Education: M.S., Electronic Information (in progress) · B.E., Cyberspace Security
Contact: yangxinmeng@tju.edu.cn