> ## Documentation Index
> Fetch the complete documentation index at: https://rootea.es/llms.txt
> Use this file to discover all available pages before exploring further.

# LLM Attacks (Prompt Injection)

> Cross-platform resources for LLM Attacks (Prompt Injection): curated HTB machines, PortSwigger labs and TryHackMe rooms with validated writeups and related skills.

# LLM Attacks (Prompt Injection)

This page aggregates **cross-platform resources to practice LLM Attacks (Prompt Injection)**: retired Hack The Box machines, PortSwigger Web Security Academy labs and TryHackMe rooms, plus curated resources (HackTricks, PortSwigger, etc.) and related skills.

## Curated resources

| Source      | Link                                                                                                 |
| ----------- | ---------------------------------------------------------------------------------------------------- |
| PortSwigger | [https://portswigger.net/web-security/llm-attacks](https://portswigger.net/web-security/llm-attacks) |

## PortSwigger labs practicing LLM Attacks (Prompt Injection) (8)

| Lab                                                                                                                                                                                                        | Difficulty                                                 | Topic       | Official                                                                                                                                                          |
| ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------- | ----------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| [Exploiting AI agents to exfiltrate sensitive information](/en/portswigger/labs/llm-attacks/ai-powered-scanner-vulnerabilities-sensitive-information-exfiltration)                                         | <span className="dbadge dbadge-easy">APPRENTICE</span>     | LLM Attacks | [Open](https://portswigger.net/web-security/llm-attacks/ai-powered-scanner-vulnerabilities/lab-sensitive-information-exfiltration)                                |
| [Exploiting AI agents to perform destructive actions](/en/portswigger/labs/llm-attacks/ai-powered-scanner-vulnerabilities-indirect-prompt-injection-via-ai-powered-scan)                                   | <span className="dbadge dbadge-easy">APPRENTICE</span>     | LLM Attacks | [Open](https://portswigger.net/web-security/llm-attacks/ai-powered-scanner-vulnerabilities/lab-indirect-prompt-injection-via-ai-powered-scan)                     |
| [Exploiting insecure output handling in LLMs](/en/portswigger/labs/llm-attacks/exploiting-insecure-output-handling-in-llms)                                                                                | <span className="dbadge dbadge-easy">APPRENTICE</span>     | LLM Attacks | [Open](https://portswigger.net/web-security/llm-attacks/lab-exploiting-insecure-output-handling-in-llms)                                                          |
| [Exploiting LLM APIs with excessive agency](/en/portswigger/labs/llm-attacks/exploiting-llm-apis-with-excessive-agency)                                                                                    | <span className="dbadge dbadge-easy">APPRENTICE</span>     | LLM Attacks | [Open](https://portswigger.net/web-security/llm-attacks/lab-exploiting-llm-apis-with-excessive-agency)                                                            |
| [Bypassing AI scanner defenses to exfiltrate sensitive information](/en/portswigger/labs/llm-attacks/ai-powered-scanner-vulnerabilities-bypassing-ai-scanner-defenses-to-exfiltrate-sensitive-information) | <span className="dbadge dbadge-medium">PRACTITIONER</span> | LLM Attacks | [Open](https://portswigger.net/web-security/llm-attacks/ai-powered-scanner-vulnerabilities/lab-bypassing-ai-scanner-defenses-to-exfiltrate-sensitive-information) |
| [Exploiting AI agents to trigger secondary vulnerabilities](/en/portswigger/labs/llm-attacks/ai-powered-scanner-vulnerabilities-exploiting-target-website-vulnerabilities-to-bypass-restrictions)          | <span className="dbadge dbadge-medium">PRACTITIONER</span> | LLM Attacks | [Open](https://portswigger.net/web-security/llm-attacks/ai-powered-scanner-vulnerabilities/lab-exploiting-target-website-vulnerabilities-to-bypass-restrictions)  |
| [Exploiting vulnerabilities in LLM APIs](/en/portswigger/labs/llm-attacks/exploiting-vulnerabilities-in-llm-apis)                                                                                          | <span className="dbadge dbadge-medium">PRACTITIONER</span> | LLM Attacks | [Open](https://portswigger.net/web-security/llm-attacks/lab-exploiting-vulnerabilities-in-llm-apis)                                                               |
| [Indirect prompt injection](/en/portswigger/labs/llm-attacks/indirect-prompt-injection)                                                                                                                    | <span className="dbadge dbadge-medium">PRACTITIONER</span> | LLM Attacks | [Open](https://portswigger.net/web-security/llm-attacks/lab-indirect-prompt-injection)                                                                            |

## TryHackMe rooms practicing LLM Attacks (Prompt Injection) (2)

| Room                                                          | Difficulty                                       | Type        | Access  | Official                                             |
| ------------------------------------------------------------- | ------------------------------------------------ | ----------- | ------- | ---------------------------------------------------- |
| [Advent of Cyber 2024](/en/tryhackme/rooms/adventofcyber2024) | <span className="dbadge dbadge-easy">EASY</span> | Walkthrough | 🟢 Free | [Open](https://tryhackme.com/room/adventofcyber2024) |
| [CupidBot](/en/tryhackme/rooms/lafb2026e6)                    | <span className="dbadge dbadge-easy">EASY</span> | Challenge   | 🟢 Free | [Open](https://tryhackme.com/room/lafb2026e6)        |

***

← [Back to the full glossary](/en/glossary)

<script type="application/ld+json">
  {`{"@context":"https://schema.org","@type":"DefinedTerm","name":"LLM Attacks (Prompt Injection)","termCode":"llm-attacks","inLanguage":"en","url":"https://rootea.es/en/skills/llm-attacks","inDefinedTermSet":{"@type":"DefinedTermSet","name":"Tactical pentesting glossary","url":"https://rootea.es/en/glossary"}}`}
</script>

*Last updated: 2026-05-09*