Oxide is pointing Claude Mythos 5 at its own firmware
Oxide announced on its blog that it has joined Project Glasswing, Anthropic’s collaborative effort to harden critical software. The company is using Claude Mythos 5 to find and fix potential vulnerabilities across its own codebase: firmware, the operating system, the control plane, and the network stack.
Co-founder and CTO Bryan Cantrill tied the model’s usefulness to how Oxide builds: the whole stack is designed from scratch, and the company controls every layer from the root of trust up through the network.
Oxide open-sources everything it writes and publishes its security findings, including a previously disclosed flaw in the TrustZone implementation on NXP’s LPC55S69 microcontroller. Its approach to working with large language models is documented in RFD 576.
Up next
Anthropic: delete the “double-check your answer” lines from your Opus 5 prompts
Anthropic’s Opus 5 prompting guide says to strip “add a final verification step” and “re-check your answer” from old prompts. Opus 5 checks its own work, so those lines just burn tokens with no quality gain. Opus 4.8 prompts otherwise carry over unchanged.
Related stories
- Rubrik found more bugs than its engineers could fix
- Anthropic adds real-time cybersecurity filtering to Opus and Sonnet
- Apiiro finds Claude Code secrets in 1,122 GitHub files
- Email leak stopgap in Claude Code checks only shell calls
- Claude Code agent sat on a Pentagon network for two months
- Irregular CEO blames human error for AI agent breaches
Comments
No comments yet. Be the first.
Join the conversation
Sign in to leave a comment. It appears after moderation.