Prompt Injection via GitHub Patch in Brave AI Chat (Leo)
HackerOneBrave Software2025-08-22
Web, AI
High
BUG BOUNTY WRITEUPS
Prompt injection and AI writeups
Prompt injection and other bugs in LLM features: data leaks through assistants, tool abuse and jailbreaks with real impact.
18 reports and writeups from 2024 to 2025. 2 carry a severity from the disclosing platform, 1 of them rated High or Critical. Last updated 2026-10-30.
Latest Prompt injection and AI reports
Search and filter all writeupsUnmasking Harmful Content in a Medical Chatbot: A Red Team Perspective
Researcher blog2024-09-05
Web, AI
Not rated
Microsoft Copilot: From Prompt Injection to Exfiltration of Personal Information
Researcher blogGitHub (Copilot)2024-08-26
Web, AI
Not rated
Google AI Studio: LLM-Powered Data Exfiltration Hits Again! Quickly Fixed.
Researcher blogGoogle (AI Studio)2024-08-21
Web, AI
Not rated
AI Under Siege: Discovering and Exploiting Vulnerabilities
Researcher blog2024-08-05
Web, AI
Not rated
Jailbreak of Meta AI (Llama -3.1) revealing configuration details
MediumMeta / Facebook (Llama)2024-07-27
Web, AI
Not rated
Zeroday on Github Copilot
Researcher blogGitHub (Copilot)2024-07-27
Web, AI
Not rated
Sorry, ChatGPT Is Under Maintenance: Persistent Denial of Service through Prompt Injection and Memory Attacks
Researcher blogOpenAI (ChatGPT)2024-07-08
Web, AI
Not rated
When Prompts Go Rogue: Analyzing a Prompt Injection Code Execution in Vanna.AI
Researcher blogVanna.ai2024-06-27
Web, AI
Not rated
GitHub Copilot Chat: From Prompt Injection to Data Exfiltration
Researcher blogGitHub (Copilot Chat)2024-06-14
Web, AI
Not rated
LLM03: Training Data Poisoning via ASCII decoding
HackerOneHackerOne2024-05-28
Web, AI
Not rated
LLM01: Invisible Prompt Injection
HackerOneHackerOne2024-05-13
Web, AI
Medium
LLM Pentest: Leveraging Agent Integration For RCE
Researcher blog2024-05-06
Web, AI
Not rated
CVE-2024-20356: Jailbreaking a Cisco appliance to run DOOM
Researcher blogCisco2024-04-18
Web
Not rated
Google AI Studio Data Exfiltration via Prompt Injection - Possible Regression and Fix
Researcher blogGoogle (AI Studio)2024-04-07
Web, AI
Not rated
New Google Gemini Vulnerability Enabling Profound Misuse
Researcher blogGoogle (Gemini)2024-03-12
Web, AI
Not rated
Bypass instructions to manipulate Google Bard AI (Conversational generative AI chatbot) to reveal its security vulnerability i.e. configuration file details exposure
MediumGoogle (Bard)2024-01-23
Web, AI
Not rated
AWS Fixes Data Exfiltration Attack Angle in Amazon Q for Business
Researcher blogAWS2024-01-18
Web, Cloud, AI
Not rated
Every entry links to the original report or article. Secure Hyperloop doesn't copy the content. Severity comes from the disclosing platform (HackerOne). Writeups without a platform rating show "Not rated". The index has 2,810 items in total.