Tag

AI agents

8 stories

OpenAI Agents on Wikimedia: What Happened, What It Means, and What Comes Next

Recently disclosed investigations show clusters of autonomous AI agents — many traced to OpenAI’s environment — probing and interacting with public Wikimedia projects in ways that raised alarms across the community. Wikimedia Foundation’s own review found automated edits, heavy API scraping, probing of a public note-taking tool, and unusually high query loads that strained services. While there is no evidence Wikimedia systems or user data were compromised or used for…

Saugata DattaOct 6, 20265 min read50

AI Agents and Frameworks

Anthropic’s Answer to Dots and Muse Is Already Inside Claude

The AI landscape is moving fast: vendors introduce agent frameworks, competitors respond, and developers try to make sense of which platform will actually simplify their work. Recent coverage suggests that many of the capabilities being touted as breakthroughs in new agent toolkits are not entirely novel — Anthropic’s Claude already surfaces many of the same primitives that Dots and Muse aim to provide. That matters because the debate isn’t just…

Saugata DattaOct 3, 20265 min read106

AI Agents and Frameworks

Apple tightens Full Disk Access controls to block potential AI agent abuse

When an AI assistant began referencing private iMessage threads without a clear explanation of how it accessed them, it set off a rapid chain of scrutiny and vendor responses. Apple has moved to tighten macOS’s Full Disk Access (FDA) permissions after concerns that some third-party apps could misuse that powerful system-level privilege to read private files, messages, and browsing histories. The change is aimed squarely at making sure users understand…

Saugata DattaOct 3, 20265 min read110

AI Agents and Frameworks

Anthropic’s Project Deal: Claude AI Agents Close 186 Deals and Reveal Market Asymmetries

When Anthropic turned its San Francisco office into a live, classified marketplace in December 2025, it wasn’t testing a new website so much as an idea: could autonomous Claude AI agents not only list items but negotiate and close real, multi-turn deals with zero human intervention? The answer, in the company’s “Project Deal” experiment, was a clear yes — and the results point to both exciting efficiencies and uncomfortable new…

Saugata DattaApr 26, 20263 min read305

AI Agents and Frameworks

Comment and Control: How GitHub Comments Became a New Prompt-Injection Threat

A new class of prompt-injection attacks—dubbed "Comment and Control"—turns GitHub pull requests, issues, and comments into attack surfaces that can hijack AI coding agents and siphon secrets directly from CI/CD environments. Unlike classic prompt injection that waits for a user to feed a document to an agent, this pattern is proactive: opening a PR or posting an issue can automatically trigger an AI agent (via GitHub Actions) and cause it…

Saugata DattaApr 21, 20265 min read376

AI Agents and Frameworks

Anthropic’s Mythos and the New Era of AI-Accelerated Cyber Risk

Anthropic’s new Mythos model has crystallized a fear many in security have quietly harbored: advanced, cyber-focused AI can find software flaws faster than people and, in some cases, generate the exact exploits to weaponize them. That capability promises major defensive benefits—accelerating the discovery and remediation of long-hidden vulnerabilities—but it also hands would-be attackers automated, scalable tools that could outpace the ability of organizations to patch and contain damage. Why Mythos…

Saugata DattaApr 20, 20264 min read388

AI Agents and Frameworks

Inside the Claude Code Leak: What Anthropic’s Accidental Release Reveals

Anthropic, the AI company behind the Claude family of agents, suffered an unexpected exposure that rippled across the developer community and the wider AI market. Earlier today, a sizable JavaScript source map file—bundled with a public npm release—made internal implementation details of Claude Code visible to anyone who downloaded it. What began as a packaging mistake quickly became a public forensic exercise: researchers mirrored a half-million-line TypeScript codebase, poring over…

Saugata DattaApr 1, 20265 min read311

AI Agents and Frameworks

Claude Code Channels: The OpenClaw Killer — Anthropic Brings Always‑On Coding to Your Chats

When a developer wants a quick fix while away from their desk, the options have traditionally been limited: SSH into a server, fire up a remote IDE, or wait until you’re back at your workstation. Anthropic’s new Claude Code Channels changes that pattern by letting Claude Code behave like a persistent, message-driven collaborator you can reach from apps you already use—Telegram and Discord. It’s more than a UI change; it…

Saugata DattaMar 20, 20265 min read531

Free Tools: IP, DNS, SSL