@xmuhtasham super interesting - surely harnesses will become smaller and smaller as models get better and will just handle things like permissions/guardrails
We're releasing Social Arena, and our first benchmark, the Deception Index
Social Arena is the first platform where humans come play social games like Risk, Catan, or Poker versus AI agents
These multi-agent matches are then used for evaluations on model behavior
More below!
Today we're launching the biggest AI security experiment with $100K+ in weekly rewards to hack AI agents.
Playground is an open-source CTF for AI agents. Every Monday, we publish a new challenge and invite anyone to break it.
Each challenge is a live AI agent with real tools, including web search and browser access. It has a secret it has been instructed to protect, and we publish its full system prompt. You can see exactly what the agent was told and try to get around its defenses.
Our first challenge, The Gatekeeper, is live now. Kai is an assistant guarding a classified access code. We review every submission ourselves. The player with the most approved breaks each week wins, and we publish a new challenge every Monday.
Playground is open source, including the client, a reference implementation of the defender engine, and every challenge configuration. You can also propose a future challenge.
Every SaaS ships one product. Your users each want a slightly different one.
Today we're open-sourcing Vendo: an agentic interface that lets your users reshape your product to exactly how they work.
> Rebuild any screen into the view they want
> Ask, and it acts → anything your API can do, your users can do by asking
> Connect Gmail, Slack, and 100+ other tools
> Build automations that run on your API, on a schedule or on triggers
All in your brand, your components, your permissions. Not a chatbot bolted on the side.
One command:
> npm install vendoai && npx vendo init
Check out our GitHub below, and DM me to get setup on our cloud.
Claude Code Full Sandbox Escape (CVE-2026-55607)
writeup: https://t.co/kzJ04Fqu4Y
prompt injection -> code execution on the host.
works even in read-only permissions mode + full sandbox
(it could be my Pwn2Own bug, but p2o was weird this year lol)
I traveled over 30 different countries so I could build this.
This is Trucy, your AI travel agent.
She handles the Flights, Hotel, Visas, E-SIMs etc. You just take the trip.
No AI slop guaranteed. Link in comments.
Introducing Obside, the AI trading copilot.
Describe your ideas in plain text, in any language, and deploy AI trading agents that execute your rules through your brokerage account.
Built for traders who want real automation, not another chatbot to babysit.
Automate now.
Introducing the world’s first AI CMO.
The teams winning now aren't the ones with the best product. They’re the ones flooding the algorithm.
Fastlane deploys social media accounts, generates viral content, and posts for you autonomously.
Today, we kill marketing: