CacheTrap
ml-security-research-lab/cachetrapResearch code for a gray-box trojan attack that flips a single KV-cache bit in fine-tuned LLM classifiers and measures per-class attack success rate.
adversarial-attackai-securitymachine-learning+2
4
ml-security-research-lab/cachetrapResearch code for a gray-box trojan attack that flips a single KV-cache bit in fine-tuned LLM classifiers and measures per-class attack success rate.
deep-jlu/skillpoisonResearch pipeline that constructs verified successful experiences and organizes them into formation records to progressively poison model skills via…
canonicalmg/reverse-captcha-evalEvaluation framework that tests whether large language models follow invisible Unicode-encoded instructions embedded in normal-looking text, with…
clearlotus-git/cve-2024-37054-pocProof-of-concept and research repository for CVE-2024-37054, an unsafe deserialization flaw in MLflow PyFunc model loading that can lead to remote…
stardust457/sldrDefense framework that mitigates malicious fine-tuning of LLMs using selective layer recovery and dynamic routing, with training, harmfulness…
azzzzliao/aegis-audio-defenseDefense framework that keeps audio-language models frozen and adds a mid-layer risk gate with late-layer safety adapters to block audio jailbreaks at…
liuz233/hardeResearch code for HARDE, an agent harness that probes and adaptively optimizes components for runtime risk detection and execution control across…
phkhanhtrinh23/agasLLM-driven agentic group shilling attack framework that manipulates black-box collaborative-filtering recommender rankings using adaptive multi-role…
binzhwang/actguardPre-execution action-auditing defense that detects and masks indirect prompt injection in tool-using LLM agents using embedding retrieval and…
yxf203/defense-as-skillRuntime guard skill that detects and blocks malicious skill-induced attacks in skill-augmented agents, with an MCTS loop that evolves the guard…
uiuc-kang-lab/persistbdResearch code for training adapters that make LLM agent backdoors survive benign fine-tuning, with trigger-rate scoring, SWE-bench evaluation, and…
chanwoochoi316/kidnapragBlack-box attack framework that hijacks reasoning in agentic retrieval-augmented generation systems by injecting poisoned documents, with support for…
zjuwyh/shadowmemDefensive framework that maintains a safety-focused shadow memory to detect and block prompt-injection and long-horizon threats against LLM agents…
alex66366/odpureBlack-box input-stage purification defense that neutralizes backdoor attacks on object detectors via corruption, diffusion reconstruction, and DBSCAN…
wsu-cyber-security-lab-ai/clipguardAdversarial image perturbation tool that uses SAM segmentation and CLIP models to evade AI-based scam image classifiers for security research.
danigil/neupermZero-retraining sanitization that reorders permutation-equivalent neural network units to destroy malware embedded via weight-steganography attacks…
eclipsemanic/threatsentry-aiThreatSentry AI is an intelligent threat hunting dashboard that leverages machine learning to proactively identify and prioritize risks in your…
shoucheng3/apache__uima-uimaj_cve-2022-32287_3-3-0Java SDK for building analytics pipelines that process and enrich unstructured text data with annotations, type systems, and modular analysis engines.