No red flags found in any of the 11 categories — no credential harvesting, no data exfiltration, no curl-pipe-shell installer. Scanned against the SlowMist agent-security taxonomy, refreshed every 8 hours. Full audit →
by AI-secure · Agent Tool · ★ 235
Last updated: · Indexed by AgentSkillsHub · Auto-synced every 8h
🔒 Is AgentPoison safe to install? View the security audit →
AgentPoison: Red-teaming LLM Agents via Memory or Knowledge Base Backdoor Poisoning 🔥🔥 Recent news please check Project page ! This repository provides the official PyTorch implementation of the following paper: [AgentPoison: Red-teaming LLM Agents via Memory or Knowledge Base Backdoor Poisoning]() Zhaorun Chen1, Zhen Xiang2, Chaowei Xiao 3, Dawn Song 4, Bo Li1,2 1University of Chicago, 2University of Illinois, Urbana-Champaign 3University of Wisconsin, Madison, 4University of California, Berkeley <img src="assets/method.png" width="
| Stars | 235 |
| Forks | 33 |
| Language | Python |
| Category | Agent Tool |
| License | MIT |
| Quality Score | 64.7098412233273/100 |
| Open Issues | 3 |
| Last Updated | 2026-08-16 |
| Created | 2024-03-22 |
| Platforms | python |
| Est. Tokens | ~16k |
Looking for a AgentPoison alternative? If you're comparing AgentPoison with other agent tool tools, these 6 projects are the closest alternatives on Agent Skills Hub — ranked by topic overlap, star count, and community traction.
Roy: A lightweight, model-agnostic framework for crafting advanced multi-agent systems using large language mo
A LLM powered agent for bioinformatics 🤖
AI Lawyer is an intelligent reasoning legal assistant powered by DeepSeek , Ollama RAG and LangChain, designed
Low code tool to rapidly build and coordinate multi-agent teams
Ship RAG based LLM web apps in seconds.
A curated list of OpenClaw resources, tools, skills, tutorials & articles. OpenClaw (formerly Moltbot / Clawdb
Explore other popular agent tool tools:
AgentPoison is [NeurIPS 2024] Official implementation for "AgentPoison: Red-teaming LLM Agents via Memory or Knowledge Base Backdoor Poisoning". It is categorized as a Agent Tool with 235 GitHub stars.
AgentPoison is primarily written in Python. It covers topics such as llm-agent, red-team, retrieval-augmented-generation.
You can find installation instructions and usage details in the AgentPoison GitHub repository at github.com/AI-secure/AgentPoison. The project has 235 stars and 33 forks, indicating an active community.
AgentPoison is released under the MIT license, making it free to use and modify according to the license terms.
The top alternatives to AgentPoison on Agent Skills Hub include Roy, BRAD, AI-Lawyer-RAG-with-Deepseek. Each offers a different approach to the same problem space — compare them side-by-side by stars, quality score, and community activity.
Grades come from a rule-based scan built on the SlowMist agent-security taxonomy, covering 11 red-flag categories including credential harvesting, data exfiltration, and curl | sh installers. It is a first-layer scan, not a manual audit — we say so rather than overstate it.
The scale of the problem is documented independently: Liu et al. (2026), in a study of 31,132 agent skills, report that 26.1% contain security vulnerabilities. Our own full-catalog census is published as a citable open dataset.
Sources & who's responsible: