Observed arrival · 2026-11-17
Kirill Novikov’s AI-security profile
A researcher’s personal site documenting work on AI control, AI-agent security, and fixes contributed to open-source projects.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- Teams assessing AI-agent and open-source security
- Worth noticing
- The homepage breaks down 27.0 billion Claude Code tokens by model and names six projects with merged fixes.
Field notes
The page links a “by model” breakdown and identifies security reviews of open-source AI-control tools as current work. It also names support from a rapid grant by BlueDot Impact and participation in Anthropic’s Cyber Verification Program; these are profile claims, not independently verified here. A linked project, Luthien, is described as an open-source AI-control proxy.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
◎NicheUnusually specific use
◉HumanPersonal, local, civic, or handmade
One card from the complete issue