AgentPoison AI Agent

by AI-secure• 250•• Scanned•Claim this listing•Request removal
Database Managementllm agentred teamretrieval augmented generation

[NeurIPS 2024] Official implementation for "AgentPoison: Red-teaming LLM Agents via Memory or Knowledge Base Backdoor Poisoning"

AgentPoison is an open-source ai agent by AI-secure, written mainly in Python, with 250 GitHub stars and the MIT license. Its code was last updated Oct 23, 2026.

Scan summary: scanned Nov 5, 2026 at commit 7b02e28. Seen able to run shell commands, run dynamic code. Scan policy

What AgentPoison can touch

  • Run shell commandsFound in the code by our scan
  • Run dynamic codeFound in the code by our scan
  • Read or write filesNot seen in this scan
  • Read environment and secretsNot seen in this scan
  • Control a web browserNot seen in this scan
  • Connect to databasesNot seen in this scan
  • Read or send messagesNot seen in this scan
  • Touch money or tradingNot seen in this scan
  • Connect to the internetNot run in this scan

"Not seen" means our checks did not find it, and the tool may still do it. Covers commit 7b02e28, Nov 5, 2026. See the scan policy.

Before you install AgentPoison

AgenticMart scanned AgentPoison on Nov 5, 2026 with automated checks for hidden code, known-malicious packages, leaked secrets and vulnerable dependencies. The result covers one commit under our scan policy, so review what access you give it and check any advisories shown here.

What it can access

  • An AI agent decides its own next steps and calls whatever tools you give it, so its risk depends on the access you grant.

Before you install

  1. Pin the version you reviewed instead of always installing the latest release.
  2. Read the tool list or scripts before the first run, and keep approval prompts on for actions that write, send or delete.
  • Maintained by AI-secure on GitHub, 250 stars and 35 forks. Popularity is not proof of safety.
  • Last code change: Oct 23, 2026.
  • Open-source license: MIT.

AgentPoison is an open-source ai agent by AI-secure, written mainly in Python. [NeurIPS 2024] Official implementation for "AgentPoison: Red-teaming LLM Agents via Memory or Knowledge Base Backdoor Poisoning"

At a glance

Stars
250
Forks
35
Language
Python
License
MIT
Created
Mar 2024
Last update
Oct 23, 2026
Type
AI Agent
Price
Free

Topics

llm-agentred-teamretrieval-augmented-generation

Questions about AgentPoison

What is AgentPoison?

[NeurIPS 2024] Official implementation for "AgentPoison: Red-teaming LLM Agents via Memory or Knowledge Base Backdoor Poisoning"

Who maintains AgentPoison?

AgentPoison is maintained by AI-secure on GitHub (AI-secure/AgentPoison). AgenticMart lists it from public repository data.

Is AgentPoison safe to use?

AgenticMart scanned AgentPoison on Nov 5, 2026 with automated checks for hidden code, known-malicious packages, leaked secrets and vulnerable dependencies. The result covers one commit under our scan policy, so review what access you give it and check any advisories shown here.

What license does AgentPoison use?

It is published under the MIT license.

AgentPoison alternatives

View more →

Similar ai agents in Database Management, ranked by GitHub stars. Compare licence, recent activity and scan status before you choose.

NameStarsLicenceLast updatedScan
AgentPoison (this page)
AI-secure
250MITOct 23, 2026Scanned, no flagged issues
Graphify
Graphify-Labs
140kApache-2.0Nov 1, 2026Scanned, no flagged issues
Claude Mem
thedotmack
108kApache-2.0Nov 1, 2026Scanned, no flagged issues
Mem0
mem0ai
68kApache-2.0Oct 30, 2026Scanned, no flagged issues
Hindsight
vectorize-io
57kMITNov 1, 2026Scanned, no flagged issues
Dokploy
Dokploy
38kNone foundOct 31, 2026Not scanned yet