Classification of Implementation-Level Security Vulnerabilities in FIDO Authentication Systems and Design Guidelines Based on Real-World Case Analysis
FIDO 인증 체계의 구현상 보안 취약점의 분류와 실제 사례 분석을 통한 설계 가이드라인 제시
FIDO holds up as a protocol, but falls over in how it is shipped. The paper narrows the threats in FIDO's own security documents down to the ones developers actually face, demonstrates them with two vulnerabilities found in Government24, and turns them into design guidelines for shipping FIDO in production.
full paper · Joohyuk Ham†, Haehyun Cho‡ · JKIISC 2026 · kci.go.kr
SCH Guard: Design and Evaluation of a Semantic Validation Core for Detecting Payload-less Attacks in LLM Agent Skills
SCH Guard: LLM 에이전트 스킬의 페이로드 없는 공격 탐지를 위한 의미론적 검증 코어의 설계 및 평가
Malicious skills can steer LLM agents toward data theft or code execution through natural-language instructions alone. This paper presents SCH Guard, a semantic validation core that combines five detection signals with an explicit decision rule. Evaluated across two LLM backbones, it improved detection accuracy with no false positives on the test set, and is available as a CLI tool for pre-merge CI checks.
short paper · Joohyuk Ham†, Haehyun Cho‡ · CISC-W'26 · github.com
IoT Smart Doorbell Vulnerability Analysis
IoT 스마트 도어벨 취약점 분석
A teardown of consumer smart doorbells — the firmware, the companion app, and the traffic between them. The team found 18 CVEs in total.
project · Team Brrester · Whitehat School · github.com