Skip to content
View euisuh's full-sized avatar
18 followers · 15 following
  • Republic of Korea Air Force
  • Seoul, South Korea
  • 07:45 (UTC +09:00)
  • LinkedIn in/euisuh
  • Instagram uiseoya

Achievements

Achievement: Pull Sharkx4Achievement: Pair Extraordinairex4Achievement: QuickdrawAchievement: YOLO

Achievements

Achievement: Pull Sharkx4Achievement: Pair Extraordinairex4Achievement: QuickdrawAchievement: YOLO

Highlights

  • Pro

Block or report euisuh

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
euisuh/README.md

Hi, I'm Euisuh 👋

Staff Sergeant · Republic of Korea Air Force (AI-Based Technology Team) · CMU CS '22
AI researcher & software engineer · Published in KOSAP & KSMI · aiXamine founding member at QCRI

Projects, publications, and writing → uiseoya.com

Pinned Loading

  1. CVEFixesUtil CVEFixesUtil Public

    Public CVEfixes dataset processor for vulnerability research and code-LLM security analysis.

    Jupyter Notebook 9 1

  2. nginx-proxy-manager-mcp nginx-proxy-manager-mcp Public

    MCP server for Nginx Proxy Manager: proxy and redirection hosts, streams, Let's Encrypt certificates (HTTP-01/DNS-01) and access lists, with dry-run previews and a read-only mode.

    Python 2 1

  3. agent-timetracker agent-timetracker Public

    Retired: a time tracker for coding agents, and a case study on why I shut it down

    Python

  4. cheoyong cheoyong Public

    Policy layer for LLM agent tool calls: allow, deny, rate-limit, escalate and audit before execution. Fail-closed; includes a Claude Code PreToolUse hook.

    Python

  5. gumiho gumiho Public

    Benchmark for indirect prompt injection in tool-using LLM agents: 24 scenarios, 5 injection channels, 6 defenses, paired benign utility

    Python