Notes
  • Welcome!
  • Windows Shells
    • Introduction
    • Command Prompt
      • Basics
      • Host Enumeration
      • Files & Directories
      • Environment Variables
      • Managing Services
      • Scheduled Tasks
      • Help
    • PowerShell
      • PowerShell vs. CMD
      • Basics
      • CmdLets & Modules
      • User & Group Management
      • Files & Dirs
      • Finding & Filtering
      • Services
      • Registry
      • Windows Event Log
      • Networking Management
      • Web Interaction
      • Scripting
      • Help
  • Windows
    • Commands
    • NTFS
  • APISEC
    • API Testing
      • Recon
      • Endpoint Analysis
      • Finding Security Misconfigurations
      • Authentication Attacks
      • Exploiting API Authorization
        • BOLA
        • BFLA
      • Improper Assets Management
      • Mass Assignment Attacks
      • SSRF
      • Injection Attacks
      • Evasion & Chaining
    • API Authentication
      • Authentication Types
      • OAuth Actors
      • OAuth Interaction Patterns
      • JSON Web Tokens
      • Claims
      • APIs & Gateways
  • PostSwigger
    • Web LLM Attacks
      • Overview
      • Exploiting LLM APIs, function, & Plugins
      • Indirect Prompt Injection
      • Leaking Sensitive Data
      • Defending Against LLM Attacks
    • JWT Attacks
      • JWTs
      • Attacks
        • Flawed Signature Verfication
        • Brute-forcing Secret Keys
        • JWT Header Parameter Injections
        • Algorithm Confusion
      • Prevention
    • OAuth
      • General Information
      • Exploiting OAuth Authentication Flaws
        • Flaws in Client Application
        • Flaws in the OAuth Service
      • OpenID
  • Red Teaming LLM Applications
    • LLM Vulnerabilities
    • Red Teaming LLMs
    • Red Teaming at Scale
    • Red Teaming LLMs with LLMs
    • Red Teaming Assessment
  • Fin
    • Course 1: Basics
      • Stocks
        • General Information
        • Shares
        • Stock Basics
      • Bonds
        • General Information
        • Components
        • Valuation
      • Markets
        • What is the Stock Market
        • What is the FED
    • Course 2: Stock Investing
  • Other
    • Learning Resources
Powered by GitBook
On this page
  1. PostSwigger
  2. Web LLM Attacks

Overview

  • Large Language Models (LLMs) are algorithms that process user inputs (via a prompt) and create a plausible response.

  • Prompt Injection is a common attack against LLMs.

  • Attacking an LLM integration is similar to SSRF: leveraging a server-side system to launch an otherwise inaccessible target, such as data and APIs.

  • Detection methodology:

    1. Identify the model's inputs, both direct (via the prompt) and indirect (training data).

    2. Figure out what data and APIs the LLM can access.

    3. Probe this new attack surface for flaws.

PreviousWeb LLM AttacksNextExploiting LLM APIs, function, & Plugins

Last updated 11 months ago