In the same amount of time it took most of us to forget our New Year’s resolutions, OpenClaw managed to upend the world of agentic AI and somehow hit 247,000 stars on GitHub in just about two months.

It’s a neck-craning trajectory. And one that left Apple and Google — who long promised the wonders of agentic AI without actually delivering them — firmly on the ground as OpenClaw began to soar.

“It took an open-source project to blow the lid off [the market],” Jentic CEO Sean Blanchfield tells Darryl K. Taft in today’s lead story.

But the fallout over security — or lack thereof — from OpenClaw created something like a pounding headache for everybody in the ecosystem. Researchers have found about 40,000 instances of AI agents that freely blab passwords when asked. In his story, Darryl writes about Jentic Mini, a new, free, self-hosted “permission firewall” that enables developers to add a layer of control around their AI agents to keep things secure.

Go deeper: OpenClaw’s biggest security flaw is why Jentic Mini exists

More stories for your Wednesday:

◉ Fivetran donates its SQLMesh data transformation framework to the Linux Foundation

◉ Claude can now open apps, click buttons, and complete tasks on your Mac — but Anthropic says risks remain

◉ HPE’s AI agents cut root cause analysis time in half

Portkey’s open-source, enterprise-grade Gateway brings governance, observability, and reliability into one layer, with an MCP Gateway for every model request and agent action. Fork it. Self-host it. Build on it.

Python is constantly evolving, and staying sharp means mastering both its modern syntax and its foundational computer science principles. These three new guides from our sister site, roadmap.sh, cover the spectrum from day-to-day scripting to high-performance logic.