Build starts limited alpha testing later this month.

1Password has launched a browser integration that lets Anthropic’s Claude use stored credentials to complete tasks on the web without the passwords ever reaching the AI model, according to a blog post published on Thursday. The company calls it a zero-exposure architecture: when Claude needs to sign in, 1Password shows the user which credential is being requested and why, then waits for biometric approval before injecting the login directly into the page. Claude never sees the vault item, password, or one-time code, and access ends when the task is complete.

The integration addresses a fundamental tension in agentic AI. Browser-based agents like Claude can navigate websites, fill out forms, and complete purchases, but reaching a login page has historically forced users to either hand over their password or take the wheel themselves. 1Password says this is the first browser integration that lets an agent use credentials without granting direct access to them.

After autofill, 1Password checks whether secrets were exposed on the page. If submission fails, the extension clears the filled values before returning control to Claude. The credential stays encrypted and controlled by 1Password throughout the process.

The 💜 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!

The launch also introduces Agentic Mode, a feature in the 1Password browser extension that automatically locks down the vault when a compatible AI agent takes control. The agent can only use logins and one-time codes explicitly approved for the current task, and the rest of the vault stays out of reach. Agentic Mode activates even if the 1Password-Claude integration is not configured, and supports agents beyond Claude.

The timing is notable given that security researchers recently demonstrated how AI browsers could be tricked into leaking user credentials through prompt injection attacks, with Anthropic’s own Claude extension among those affected. 1Password CTO Nancy Wang said in the company’s announcement that the answer is not handing agents your secrets, but letting a user give an agent permission to use a credential without letting the agent see it. She called that distinction the foundation of trust in AI agents.

1Password for Claude is available now on Mac for business, family, and individual plans, and requires the 1Password desktop app, browser extension, Claude desktop app, and Claude browser extension. The company, which recently acquired Israeli startup Apono to govern AI agent access inside enterprise systems, said it plans to add support for payment cards and identity details after launch.

CNET’s password manager expert Joe Supan said he would normally be very wary about giving an AI agent access to his password manager, but that 1Password appears to have several good guardrails in place, particularly biometric authentication for each login. The integration marks the first time a major password manager has built a dedicated secure channel for an AI agent to use credentials at runtime, rather than exposing them to the model’s context. Whether the approach holds up against the kind of prompt injection attacks that have already compromised AI browsers remains to be seen.

The brand iRobot launched the first Roomba robot vacuum back in 2002, and popularity for the handy devices skyrocketed from there. Countless competitors have emerged, but Roomba is still going strong. Its latest models have all the new features we love, from doubling as a vacuum and a mop to fantastic navigation and suction. The Roomba Max 705 is currently keeping my house clean as I test it for our robot vacuum guide, and it’s doing a great job both mopping and vacuuming the floors in my massive second story.

Looking to own your own Roomba robot vacuum? These iRobot promo codes and coupons will make it more affordable than ever to own your own powerful robot vacuum from the iconic brand.

Save $400 Off for Fathers Day at iRobot

Give Dad the gift of a clean home (that’s easier than ever) with up to $400 off bestselling iRoomba robot vacuums. This means hundreds of dollars off popular models like the Roomba Plus 505 Combo, Roomba Max 705, and more. Act soon before Father’s Day passes and the deals disappear. A clean home has never been easier (or cheaper) to achieve.

iRobot Promo: Treat a Friend to 15% Off a Robot and Save $15 on iRobot Accessories

I once gifted a friend a robot vacuum and she regularly tells me how it’s a marriage saver. If you want to change a friend’s life (and potentially save their marriage) you can gift a friend 15% off a robot vacuum with this iRobot coupon code, and you’ll get $15 off robot accessories for yourself. A clean win for everyone.

Save 33% on the Roomba Max 705 With Anti-Tangle Brushes Using This iRobot Offer

I test lots of robot vacuums, and my favorite from iRobot’s lineup is the Roomba Max 705. It’s packed with two anti-tangle brushes, excellent navigation, a self-emptying base station, and left my home sparkling clean. It’s a great choice for a Roomba for pet hair, and it’s done a great job with the litter my cat leaves all over the house, too. Extra bonus: it’s gorgeous. You can get 33% off this powerful robot vacuum right now with this iRobot discount code.

Tackle Every Mess Effortlessly: $300 Off the Roomba Plus 405 Combo Robot With AutoWash Dock

Looking for something a little more affordable? The Roomba Plus 405 Combo Robot is cheaper but still packs mopping and vacuuming, and still comes with an auto-emptying docking station. It’s already a pretty affordable model and a great iRobot vacuum for hardwood floors, and right now you can get it for $300 cheaper with this iRobot coupon.

Save $290 on Smart Cleaning: The Roomba 205 DustCompactor Combo Robot Vacuum & Mop

If you don’t want a big base station in your home but still want a robot vacuum that can vacuum and mop, the Roomba 205 DustCompactor is for you. The Roomba 205 will compact and store debris for up to 60 days inside the robot itself, no emptying station required. It’s bagless, like my favorite Shark robot vacuum, so you don’t need to worry about buying and replacing bags to use the vacuum. If that sounds up your alley, don’t miss your chance to save almost $300 on this robot vac with this iRobot promo code.

Get 10% Off Your First iRobot Purchase When You Sign up for Emails

Not sure what you want to get from iRobot, but know you want an iRobot discount code? You can sign up for iRobot’s emails and get a code for 10% off whatever your first purchase is. Enjoy!

Save $500 on Max 705 Combo

The Roomba Max 705 Combo robot + AutoWash dock is the perfect household helper. With a self-cleaning roller mop, detailed edge cleaning, anti-tangle dual rubber brushes, and a powerhouse dock, this impressive robovac-mop combo takes all of the gruntwork out of the normal household cleaning chores. Usually $1,300, the Max 705 is nearly half off at $800. If you’ve been looking to upgrade your cleaning routine, snag this model for $500 less than usual, for a limited time.

Get 20% Off iRobot Accessories

An iRobot is an investment, requiring parts replacement, upgrades, and routine maintenance to keep your device in tip-top shape. iRobot knows this and wants to keep the costs down while maintaining the longevity of this top tech. Be sure to check out iRobot Accessory Offers for huge savings on genuine iRobot accessories for your robot, including up to 20% off with bundle deals. With this “buy more, save more” sale, you can build your own bundle for increased savings, so when you buy any 2, you’ll get 10% off; buy any 3, get 15% off; or when you buy any 4, you’ll get 20% off on eligible accessories. So whatever your need is, make sure you take advantage of this deal and stock up on essential accessories like filters, dirt disposal bags, replacement parts and more.

U.S. beverage maker Coca-Cola said one of its dairy subsidiaries was hacked and that it’s shutting down its operations for the foreseeable future. The multinational giant said in a disclosure with the U.S. Securities and Exchange Commission that its Fairlife dairy company was hit by ransomware and that its production systems are affected. The company said that its Fairlife production operations across the United States are “temporarily suspended.”

Fairlife’s operations in Canada are unaffected.

Coca-Cola is one of the largest companies in the world, with products spanning carbonated drinks, water, and dairy products. Its Fairlife dairy is one of the company’s major brands, with an estimated $4 billion in sales by 2024.

Ransomware attacks on food and beverage companies can have lasting effects. Past incidents at Arizona Beverages in 2019 and food distributor giant UNFI last year resulted in weeks-long disruptions to their respective production lines and empty grocery shelves.

Coca-Cola didn’t say when Fairlife’s systems would be restored.

Do you know about the cyberattack at Fairlife? Do you work at the company? We would love to hear from you. From a non-work device, you can securely contact Zack Whittaker on the Signal messaging app with the username zackwhittaker.1337.

Weekly newsletters

Get daily news, weekly deals and the week’s top tech stories

Commenting access

Join the conversation, share your thoughts and get expert advice

Member badges

Earn badges as you explore news, deals, reviews, guides and more

Exclusive deals

Save on gadgets, subscriptions and accessories with handpicked discounts

Presented by Ping Identity

Enterprises need to treat zero trust security architecture as an immediate requirement for AI agents rather than a long-term goal, says Andre Durand, CEO and founder of Ping Identity. Zero trust, the security model built on the assumption that no user, device, or system should be automatically trusted, requires continuous verification before every action rather than a single check at login. Agentic AI has profoundly compressed the risk timeline enterprises must manage, demanding that permission decisions be evaluated in real time.

That compression shows up in how permissions accumulate. Every time an employee approves an AI agent’s request for access to a company drive, a database, or a code repository, the enterprise hands over a sliver of control that looks routine in isolation. Across thousands of agents making thousands of requests, those approvals accumulate into an exposure that most existing security architectures were never built to measure.

“The rise in desire to use agents right now, and the speed of agentic, is highlighting the need to move faster on the principles of zero trust,” Durand says. “Agents just move faster, full stop. A human compromise might be measured in minutes or hours, sometimes days. At agentic speed, a thousand actions could happen in five minutes.”

Why zero trust is now urgent for agentic AI

That difference in velocity changes how enterprises need to think about permissions. Two variables matter: the surface area of access an agent is granted and the duration that access remains valid. Traditional identity and access management tends to grant broad permissions and leave sessions open for extended periods because the human using them moves at human speed. Zero trust, in contrast, collapses both variables at once by narrowing access down to what is strictly necessary and revalidating it continuously, rather than once at login.

“Zero trust really just says, just enough, just in time,” Durand says. “It’s your next action that we care about. We’re moving identity from an era where access was our runtime control point — meaning were you logged in, did you have a session — toward the decision that sits behind that login.”

Why agents must be treated as first-class identities

That shift to decision-based control has direct implications for how agents should be provisioned in the first place. The common practice of letting an agent operate under a cloned human login or a shared service account doesn’t work, Durand says.

“Each agent should have its own identity,” he explains. “It should not be impersonating the human. It can act on behalf of the human, we could explicitly delegate authority to an agent, but we don’t want to blur the lines between the human taking action and the agent taking action.”

And beyond that is another concern: the shared secrets, API keys in particular, that many service accounts still rely on. For example, the habit of embedding keys directly in source code, where they can be committed accidentally and exposed, is a convenient but weak security pattern that agentic workflows make considerably riskier. Building service account architectures that let agents authenticate without relying on those shared credentials or other long-lived standing access is now an urgent priority rather than a long-term cleanup project.

Where enterprises can enforce zero trust policies

Enforcing any of this in practice requires identifying where policy can actually be applied. Several existing choke points, including API gateways and the agent gateway sitting in front of MCP servers, offer practical locations where enterprises can inspect what an agent is requesting and apply policy rules before granting it.

“Those policies could leverage real-time risk and fraud signals, and then enforce, deterministically, what the agent can do when it interacts with these systems,” Durand explains.

The goal is to move authorization from something decided once at login to something evaluated at the moment of every consequential action, such as an agent attempting to commit code to a repository. Instead of carrying a standing permission to write to GitHub, the agent’s request would be checked against context and policy at that specific moment, closing the window of trust down to the scope of a single action.

Stopping AI agents from rewriting their own permissions

That model becomes especially important given how agents can behave once they are already inside a system — for example, coding agents that have acknowledged, when questioned, either ignoring a specific guardrail entirely, or attempting to rewrite the permissions they were given.

“Who’s watching the watcher? Zero trust needs to apply here,” Durand says. “If generative AI systems follow your instruction 97% of the time, and you’re simply asking it for advice, that might be fine. If it’s responsible for making a decision about who gets let in, 97% is not good enough.”

How to trust AI-generated output at agent speed

The answer to that gap is not to eliminate AI from the review process, but to structure reviews so no single agent’s judgment is taken at face value. Because human review cannot scale to the volume and speed of agentic output without erasing the advantage of using agents at all, a new framework is necessary, so that when one agent produces work, such as code, separate agents evaluate it, provided those reviewing agents are kept from communicating with one another or with the one they are checking. It’s a new human-AI paradigm, Durand says.

“We probably will have to develop frameworks that we trust without seeing or verifying the output directly,” he explains. “It’s not that that construct is 100% foolproof. However, it’s the best we can do to move at agent speed. We can’t trust the exact output, but we can trust the framework.”

In practice, that means combining automated review with clear human accountability for higher-risk decisions, rather than treating agent output as self-validating.

For traditional auditors, reviewing every transaction individually is never feasible, and statistically valid sampling stands in for full verification. The same applies to risk accumulation: a single agent action might carry little risk on its own, while a sequence of actions moving in a consistent direction could cross a threshold that triggers an intervention, including a kill switch capable of halting the agent before further harm occurs.

What to ask when evaluating agentic identity platforms

For security leaders evaluating identity platforms for agentic AI, there’s no narrow checklist. Enterprises should evaluate what their full lifecycle of agent management looks like. Most enterprises are managing agents on two fronts simultaneously: customer-facing agents acting on behalf of external users, and internal agents deployed to automate enterprise processes.

“Pause long enough to see the totality of what it would mean to secure multiple agents, both interacting with you from the outside as well as being deployed on the inside,” Durand says. “We need discovery and visibility of all the agents operating within our estate, a place to register them, a standard way to assign custodians, and a way to construct and centralize policy so security can enforce it across the organization.”

And while basic security principles were already fully understood before agentic AI arrived, what has changed, Durand says, is that the cost of moving slowly has finally caught up with the cost of moving carelessly, giving enterprises a narrowing window to build the right architecture before widespread agentic adoption makes retrofitting far more expensive.

Sponsored articles are content produced by a company that is either paying for the post or has a business relationship with VentureBeat, and they’re always clearly marked. For more information, contact sales@venturebeat.com.

Keytec brought the Magic Touch to the 1994 Summer Consumer Electronics Show in Chicago. The Texas company, founded in 1987, offered a straightforward way to give standard CRT monitors and notebook screens touch input without replacing the entire display. At a moment when keyboards and mice defined personal computing, the idea of pressing a finger directly on the glass stood out as genuinely forward-looking.

The Magic Touch’s hardware took the shape of a framed overlay, similar to a plastic frame with a clear adhesive membrane inside. This membrane rested on top of whatever monitor you were using at the moment, with a nice border that matched the forum’s standard beige or black color scheme. Within this jumble of layers was a brilliant innovation: an invisible spacer system that kept the whole device light and didn’t interfere with your view while also protecting the screen from scratches. The membrane handled roughly 80% of the display, which isn’t awful, and it was strong enough to withstand the ordinary 3H pencil scratch.

Sale

InnoView 15.6″ Portable Monitor 60Hz Touchscreen 1080P 10-Point Touch Screen Monitor Portable with…

• [10-Point Touchscreen Portable Monitor]: Portable screen compatible with Windows and MacOS systems. You can get touch function for your laptop by…
• [Get a Monitor Protective Sleeve]: The case is tailor-made for your portable laptop monitor, lightweight and durable, easy to carry, a perfect…
• [FHD IPS Portable Display]: 15.6 inch 1080P portable screen for laptop adopts a real reliable IPS screen with a viewing angle of 178°. Compared with…

To make things work, you needed a little external controller. Early versions were connected via an obsolete serial port, whereas subsequent versions just used USB. The panel simply hooked into the controller, and many configurations provided power via the computer connection, eliminating the need for a bulky power brick on your desk. The items were offered in a number of sizes, ranging from 12 to 17 inches for poor laptops to 13 to 24 inches for desktop displays, with larger cousins appearing in related products. It was very simple to install; simply attach some clips or brackets for large desktop monitors that would clip over the top of the bezel, or use adjustable straps for laptops to keep everything in line with the screen. You finished in a few minutes, and the old screen was as good as new underneath.

Once everything was set up and connected, you’d need some software to instruct the touchpad what to do. This would convert all of the touchy feely inputs into mouse operations. You’d go through a simple calibration process to get the pixels and the membrane in line. Then you could choose whether to click on contact or lift off to get the device to work, and you could even change for left or right hand preference, as well as temporary right click on using a software toggle. It performed all of the standard mouse actions: cursor movement, single clicks, double taps, and drags. The touch resolution was great and high, 4096 by 4096 points, and the entire thing responded to finger pressure, which ranged from 50 to 120 grams per square centimeter.

The effectiveness of the system was determined by the software being used. Big buttons on interface elements functioned perfectly with a finger or a stylus, and menu navigation was simple. You’re probably familiar with some of the older action games, which were more hit and miss. Because of the small physical space between the membrane and the screen, you had to make sure you had everything set up correctly for those tiny targets; otherwise, you’d have bizarre parallax, and glare may be an issue, especially if you had a shiny display. It also took some force to work, albeit not much.

After completing the initial Magic Touch launch, the company continued to develop subsequent versions that worked with a variety of operating systems and connection methods. They made this type of device until 2017, when it was handed over to a new business that continued to create custom touch-focused solutions for a variety of monitors and panels. Nowadays, you can get very much the same idea, just updated, and still purchase a Magic Touch-style gadget to convert an old display into a touchscreen.

One features an optical illusion spoof of ‘They Live.’

Roblox announced Build on Wednesday, a new creation tab inside the Roblox mobile app that lets anyone turn a text prompt into a basic playable game without touching Roblox Studio or writing a line of code. A creator can describe something like a cozy forest adventure game with environmental obstacles, and Build will generate a starting point with gameplay mechanics, environment, characters, sound, and visual style. The feature begins public alpha testing in New Zealand on July 28 for age-verified users nine and older.

Build shares a back end with Roblox Studio, meaning creators can start a project on their phone and continue refining it on desktop with the full Studio toolset, or launch agents from Studio and check progress from mobile. The system is powered by a mix of open-source and proprietary Roblox AI models trained on what the company describes as a uniquely large set of 3D models and gaming-specific data. Roblox’s Cube foundation model, which the company introduced earlier this year alongside agentic Studio tools, generates game-ready objects that can drive, shoot, or otherwise behave as expected without manual scripting.

Roblox said it is aware of the quality risk that comes with lowering the barrier to game creation. The company said its discovery system ranks games by long-term retention, not recency or volume, and that games nobody plays will not surface on the homepage regardless of how they were made. Published games from Build will go through the same safety checks and retention-based discovery ranking as all other Roblox titles, and games targeting younger players will undergo an extended review before being added to the Roblox Kids or Select catalogues.

Alongside Build, Roblox is shipping a suite of agentic tools for professional creators over the coming months. These include a playtesting agent that finds bugs before players encounter them, an analytics agent that answers questions about game performance in plain language, and an experiment agent that identifies tests to improve engagement and monetisation. A new scene-generation model is also in development that will create entire editable and playable 3D environments from a single text prompt.

The announcement comes as AI-generated content is flooding platforms faster than review systems can handle it, a dynamic that drove an 84 percent jump in App Store submissions and prompted Apple to crack down on low-quality AI-built apps. Roblox is betting that its retention-based discovery system will serve as a natural quality filter, surfacing only games that players actually want to keep playing.

A base version of Build will be free, with paid options for power users. The rollout will expand to more regions after the New Zealand alpha, and Roblox said it plans to share further creation tool announcements in the coming months. The platform has 132 million daily active users, any one of whom could now prototype a game from their phone.