PaperTrace

Interactive paper reading
for AI security

This minimal build keeps the interactive reading structure for security papers, with focused notes, diagrams, and source links.

Web SecurityBlack Hat SEOSearch Abuse

Into the Dark: Unveiling Internal Site Search Abused for Black Hat SEO

Zhang et al. Β· USENIX Security 2024

A large-scale measurement of how internal site search pages become indexable promotion carriers for black-hat SEO.

β†’
AI SecurityCAPTCHAMultimodal LLMs

Oedipus: LLM-enchanced Reasoning CAPTCHA Solver

Deng et al. Β· ACM CCS 2025

A reasoning-CAPTCHA paper showing how direct LLM failures can be converted into a DSL-guided decomposition pipeline.

β†’
AI SecurityCAPTCHAVLM Agents

Are CAPTCHAs Still Bot-hard? Generalized Visual CAPTCHA Solving with Agentic Vision Language Model

Teoh et al. Β· USENIX Security 2025

A VLM-agent paper showing how visual CAPTCHA solving can be reframed as abstraction plus search over interactive UI states.

β†’
AI SecurityPrompt InjectionLLM Agents

Not what you've signed up for: Compromising Real-World LLM-Integrated Applications with Indirect Prompt Injection

Greshake et al. Β· AISec '23 2023

The foundational indirect prompt-injection paper: adversarial instructions planted in retrieved data can steer LLM apps and tool calls.

β†’