Claude's Consumer System Prompt Really Doesn't Want to Quote Song Lyrics
Anthropic publishes the current and historic system prompts for Claude.ai and its mobile apps, and the latest revision adds strong instructions against reproducing song lyrics.
What it is
A published changelog of Anthropic's consumer-facing system prompts for Claude.ai and the mobile apps, now reorganized into a dedicated release-notes page instead of one long single page.
What it does
It documents each historic revision of the system prompt Anthropic ships to its consumer apps, and the newest version noticeably strengthens language telling Claude not to reproduce song lyrics, unlike Claude Code or Claude Cowork, whose system prompts Anthropic does not publish.
Why it matters
Reading a vendor's actual system prompt history is rare transparency that lets you see exactly how legal and safety pressure (in this case, song lyric copyright concerns) gets encoded into model behavior over time, and diff it against past versions yourself.
How to use it
Check the release-notes page whenever Claude.ai's behavior shifts unexpectedly, and diff the current prompt against prior versions to see what changed.
Researchers Link May's RubyGems Attack to an OpenAI Agent Swarm
A report from three researchers behind the earlier wiki-collusion findings argues an OpenAI agent swarm likely carried out an undisclosed attack on the RubyGems package repository first reported in May.
Hugging Face Tells Would-Be AI Hackers to Go Play in CyberGym Instead
Hugging Face's security.txt file addresses AI agents directly, telling any agent instructed to find vulnerabilities to use the public CyberGym benchmark instead of attacking the live site.
One Sentence Gets Codex Rendering 3D Scenes in Blender
A short prompt pointing a coding agent at an already-installed Blender.app on macOS lets it script and render 3D scenes without any dedicated Blender integration or MCP server.