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
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.
xiaoyan-lisa/defenses-for-tool-integrated-llmResearch code and experiments for defending tool-integrated LLM agents against adversarial attacks, extending Agent Security Bench with new defense…
zhou-cybersecurity-ai/synghostSyntactic Ghost: An Imperceptible General-purpose Backdoor Attacks on Pre-trained Language Models
siyuanli00/intactResearch code reproducing multi-turn LLM jailbreak experiments (FITD, MRCJ, ActorAttack, X-Teaming) from the SoK intent-oriented systematization…
zdcslab/rubrics-as-an-attack-surfaceResearch code for Rubric-Induced Preference Drift (RIPD): evolutionary rubric search, benchmark-preserving selection, and DPO policy misalignment…
96kihyun/batch_jailbreakThis repository contains the official implementation of the paper "[Safety in Batches? Understanding and Mitigating Safety Failures in Batch…
s3n4t0r-0x0/bear-c2The exploit server for out-of-band findings. Point a target at a domain you own. Every HTTP request and every email it sends back lands in a…
phkhanhtrinh23/agasLLM-driven agentic group shilling attack framework that manipulates black-box collaborative-filtering recommender rankings using adaptive multi-role…
ghostwolflab/ai_security_topA structured knowledge base covering AI security fundamentals, threat modeling, red team offensive techniques, and blue team defenses, including LLM…
atsukisato/pgm-attackResearch code for poisoning attacks on the PGM-index, demonstrating how to craft adversarial data to degrade learned index performance.
pwnhyo/t-mapTrajectory-aware evolutionary search framework for red-teaming LLM agents over MCP servers, generating adversarial prompts to map vulnerability…
chanwoochoi316/kidnapragBlack-box attack framework that hijacks reasoning in agentic retrieval-augmented generation systems by injecting poisoned documents, with support for…
jaewonalive/camodocsResearch implementation of a poisoning attack against retrieval-augmented language models using camouflaged documents to evade filtering defenses and…
ai45lab/skill-safety-benchBenchmark for evaluating AI agent safety against attacks embedded in skill-facing context, with 155 cases across 6 risk domains, measuring task…
davidcarliez/trustmebroBypass llm guardrails by confusing it with fabricated tool output.
george0papasotiriou/llm-jailbreak-via-chain-of-logic-injection-cve-2026-3098Multi-stage prompt injection technique that bypasses LLM safety alignment via identity reassignment, refusal suppression, and output coercion,…
pppyb/secopdResearch implementation for mitigating adaptive prompt injections via on-policy distillation, with training recipes and evaluators for SEP, PISmith,…
csoai-org/memory-poisoning-axis🧪 Measures memory-poisoning / prompt-injection deterministically — anchored to CVE-2026-24301 (CoSnitch), Inspect scorer, signed receipts.…