AI BriefingDeletion-inducing attack on Claude Code
Key point
Evidence has emerged that the TCRF site delivered file deletion instructions targeting AI agents.
Details
While investigating PSX game data, Claude Code detected a prompt injection payload on a The Cutting Room Floor (tcrf.net) page.
The payload instructed the agent to empty files in the repository and replace them with other files. Claude Code refused the instruction and warned the user. According to reported evidence, the problematic content was conditionally served when identified as an AI user agent.
The poster presented URLScan responses and GitHub analysis reports as evidence. This case illustrates the risk of AI agents using web search and browsing capabilities misinterpreting untrusted external content as execution instructions.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Australian PM Reports OpenAI Agent Hacked Medicare Portal in Believed First Known AI-Driven Government Breach