InfoSec Write-ups: CVE-2025-54794 Claude AI Prompt Injection

web

infosecwriteups.com·infosecwriteups.com/cve-2025-54794-hijacking-claude-ai-wi...

A practical security disclosure illustrating prompt injection risks in deployed LLMs; relevant to AI safety practitioners focused on adversarial robustness and secure deployment of frontier models.

Metadata

Importance: 42/100blog postanalysis

Summary

A security writeup documenting CVE-2025-54794, a prompt injection vulnerability in Claude AI that enabled jailbreaking and potential hijacking of the model's behavior. The article details how an attacker could craft malicious inputs to override Claude's safety instructions and elicit unintended responses. This serves as a concrete case study in real-world AI system exploitation via prompt injection.

Key Points

•Documents a formally assigned CVE (CVE-2025-54794) specifically for a prompt injection attack against Claude AI, indicating recognized severity.
•Demonstrates how prompt injection can bypass Claude's built-in safety guardrails and cause the model to produce restricted or harmful outputs.
•Illustrates the 'jailbreak' technique where carefully crafted user input overrides system-level instructions governing model behavior.
•Highlights the challenge of securing LLM-based systems against adversarial inputs at the application and model layer.
•Relevant to AI red-teaming and deployment safety, showing that prompt injection remains a critical unsolved vulnerability class for deployed LLMs.

Cited by 1 page

Page	Type	Quality
Anthropic	Organization	74.0

Cached Content Preview

HTTP 200Fetched Feb 25, 202616 KB

[Sitemap](https://infosecwriteups.com/sitemap/sitemap.xml)

[Open in app](https://play.google.com/store/apps/details?id=com.medium.reader&referrer=utm_source%3DmobileNavBar&source=post_page---top_nav_layout_nav-----------------------------------------)

Sign up

[Sign in](https://medium.com/m/signin?operation=login&redirect=https%3A%2F%2Finfosecwriteups.com%2Fcve-2025-54794-hijacking-claude-ai-with-a-prompt-injection-the-jailbreak-that-talked-back-d6754078b311&source=post_page---top_nav_layout_nav-----------------------global_nav------------------)

[Medium Logo](https://medium.com/?source=post_page---top_nav_layout_nav-----------------------------------------)

Get app

[Write](https://medium.com/m/signin?operation=register&redirect=https%3A%2F%2Fmedium.com%2Fnew-story&source=---top_nav_layout_nav-----------------------new_post_topnav------------------)

[Search](https://medium.com/search?source=post_page---top_nav_layout_nav-----------------------------------------)

Sign up

[Sign in](https://medium.com/m/signin?operation=login&redirect=https%3A%2F%2Finfosecwriteups.com%2Fcve-2025-54794-hijacking-claude-ai-with-a-prompt-injection-the-jailbreak-that-talked-back-d6754078b311&source=post_page---top_nav_layout_nav-----------------------global_nav------------------)

![](https://miro.medium.com/v2/resize:fill:64:64/1*dmbNkD5D-u45r44go_cf0g.png)

[**InfoSec Write-ups**](https://infosecwriteups.com/?source=post_page---publication_nav-7b722bfd1b8d-d6754078b311---------------------------------------)

·

Follow publication

[![InfoSec Write-ups](https://miro.medium.com/v2/resize:fill:76:76/1*SWJxYWGZzgmBP1D0Qg_3zQ.png)](https://infosecwriteups.com/?source=post_page---post_publication_sidebar-7b722bfd1b8d-d6754078b311---------------------------------------)

A collection of write-ups from the best hackers in the world on topics ranging from bug bounties and CTFs to vulnhub machines, hardware challenges and real life encounters. Subscribe to our weekly newsletter for the coolest infosec updates: [https://weekly.infosecwriteups.com/](https://weekly.infosecwriteups.com/)

Follow publication

# 🧠 CVE-2025–54794: Hijacking Claude AI with a Prompt Injection — The Jailbreak That Talked Back

## A high-severity prompt injection flaw in Claude AI proves that even the smartest language models can be turned into weapons — all with a few lines of code.

[![Aditya Bhatt](https://miro.medium.com/v2/resize:fill:64:64/1*6TFmlC58KtmaRsYHdjy9Qg.jpeg)](https://medium.com/@adityabhatt3010?source=post_page---byline--d6754078b311---------------------------------------)

[Aditya Bhatt](https://medium.com/@adityabhatt3010?source=post_page---byline--d6754078b311---------------------------------------)

Follow

4 min read

·

Aug 6, 2025

14

1

[Listen](https://medium.com/m/signin?actionUrl=https%3A%2F%2Fmedium.com%2Fplans%3Fdimension%3Dpost_audio_button%26postId%3Dd6754078b311&operation=register&redirect=https%3A%2F%2Finfosecwriteups.com%2Fcve-2025-54794-hijacking-claude-ai-w

... (truncated, 16 KB total)

Resource ID: ab73eadb1eca528e | Stable ID: MDgyYjdmYW