The Register argues that AI coding agents behave like software that will ingest untrusted instructions, connecting the jqwik maintainer’s anti-AI output warnings with Shai-Hulud-style supply-chain attacks and the broader risk that bots can be manipulated through open-source project text and build artifacts.
AI is code – and can't be prompted into being smarter
The Register argues that AI coding agents behave like software that will ingest untrusted instructions, connecting the jqwik maintainer's anti-AI output warnings with Shai-Hulud-style supply-chain attacks and the broader risk that bots can be manipulated through open-source project text and build artifacts.
Source: The Register