Prompt Security | From SentinelOne

Blog　·　Research

ChatGPT Security Guide: Enterprise Risks, Incidents, and Practitioner Guidance

Explore Research

Blog　·　Research

The Agentic AI Attack Surface: Where Risk Lives Beyond the Prompt

Explore Research

Blog　·　Research

From Trivy to LiteLLM: Expanding the LLM Supply Chain Threat Model

Explore Research

RESOURCES　·　Updated weekly

One Place for Every AI Security Resource.

News, research, tools, and analysis - everything a security team needs to understand and respond to the AI threat landscape.

Tools

Open-source tools from Prompt Security

ClawSec

Secure AI agent skills

Security skill suite for AI agents. Drift detection, live recommendations, automated audits, and skill integrity verification.

View on Github

OneClaw

OpenClaw visibility and oversight

Track and analyze OpenClaw deployments across your organization. Visibility into agent activity, usage patterns, and security insights.

View the Tool

ps-fuzz

Test system prompt resilience

Test and harden your system prompt against dynamic LLM-based attacks. Supports 16 providers and 16 attack types.

View on Github

Start Here

Visual Prompt Injection

In a visual prompt injection, malicious instructions are hidden inside an image rather than the text prompt, designed to be read by the model but invisible to the person looking at it.

Start Here

Toxic, Biased or Harmful AI-generated Content

An LLM, jailbroken or not, can produce toxic, biased, or otherwise harmful output. Because model behavior is non-deterministic, this risk exists even when a system is working exactly as designed.

Start Here

Shadow AI

Shadow AI describes the AI tools and agents employees adopt on their own, from chat assistants to coding copilots to autonomous agents, without visibility or approval from IT or security teams.