Claude Code Permission Modes in 2026: What `--allowedTools`, Whitelists, and Sandbox Boundaries Actually Restrict
Claude Code Permission Modes in 2026: What , Whitelists, and Sandbox Boundaries Actually Restrict This article was written with the assistance of AI, under human supervision and review. Most Claude Code security failures stem from treating permission modes and sandbox boundaries as interchangeable concepts. Teams ship agents with Auto mode enabled, assuming the Bash sandbox prevents harm. The sandbox restricts file system access. Permission modes control whether Claude even attempts the action. When these two layers misalign, the agent either breaks production workflows or leaks credentials through unrestricted tool calls. The confusion compounds when engineers discover and assume whitelisting individual capabilities (like or ) provides complete protection. It does not. The whitelist...
The VMs Powering Mobile Agents (Instinct, Claude Code)
Article URL: https://rohanadwankar.github.io/posts/platforms.html Comments URL: https://news.ycombinator.com/item?id=49605644 Points: 13 # Comments: 1
Cloudflare Spectrum for Minecraft: Protect, Accelerate, Then Monitor with Crystade
Protecting and Monitoring Minecraft Servers: Cloudflare Spectrum & Crystade Minecraft servers, especially public networks, face constant threats: DDoS attacks, network congestion, and silent failures that drive players away. Cloudflare Spectrum has become a popular solution for protecting non-web traffic like Minecraft’s TCP/UDP protocols. But protection alone isn’t enough—you also need to know when your server is down or misbehaving. This article explores what Spectrum does well, where it falls short, and how Crystade completes your infrastructure. What Cloudflare Spectrum Does for Minecraft Cloudflare Spectrum acts as a proxy for any TCP or UDP-based application, including Minecraft. Instead of exposing your origin server directly to the internet, traffic hits Cloudflare’s global edge...
Coop – Isolated VM Environments for Running Claude Code and Codex
Article URL: https://github.com/trailofbits/coop Comments URL: https://news.ycombinator.com/item?id=49593842 Points: 7 # Comments: 1
Boox’s tiny Picco e-reader should land in November
Boox teased the Picco, its take on the buzzy Xteink X4 e-reader, back in July, but provided almost no details. Now, thanks to some reporting out of IFA, we've got a bit more info, though sadly still no price. Unlike the Xteink devices, the Picco will not have magnets on the back for attaching to a phone. The 3.97-inch e-paper display, however, will be touchscreen, which The Verge's Andrew Liszewski called a "welcome upgrade" on the X4 Pro. The Picco has physical page-turn buttons, and they can be customized with different actions for short and long presses. There's apparently no onboard storage, and it relies entirely on a microSD card to h … Read the full story at The Verge.
Research acceleration: The view inside OpenAI
Article URL: https://openai.com/index/research-acceleration-view-inside-openai Comments URL: https://news.ycombinator.com/item?id=49587217 Points: 46 # Comments: 18
Slackware 16 Alpha 1 Available Following GCC 16.2, Glibc 2.44 & Binutils 2.47 Upgrade
The timing on Saturday's Slackware-based Zenwalk ISO "current milestone" update is now more clear with upstream Slackware having prepared what is its first alpha release of the upcoming Slackware 16...
Cloud in a Bottle: making self-hosting accessible to everyone
Article URL: https://cloudinabottle.org/blog/launch-post Comments URL: https://news.ycombinator.com/item?id=49582000 Points: 33 # Comments: 10
Remote Access to Termux via SSH: Managing Your Smartphone from a PC
Managing files, running scripts, and executing terminal commands on your smartphone from a computer is much more convenient through a full SSH session, eliminating the need to switch between device screens. Installing OpenSSH Open Termux and install the package for the secure protocol: Configuration Set a password to secure the connection: Find your current username using the command (usually looks like ). Start the SSH server: Connecting from a PC Find your smartphone's local IP address using (look for the line under the interface). Note that Termux uses a non-standard port 8022 by default instead of the standard 22. Connect from your computer (Linux, macOS, or Windows via PowerShell/PuTTY): After entering your password, you will gain full access to your phone's terminal directly...
Portal by Spotify cut my Claude Code token usage by 90%
Article URL: https://engineering.atspotify.com/2026/9/portal-by-spotify-cut-my-claude-code-token-usage-by-90 Comments URL: https://news.ycombinator.com/item?id=49571465 Points: 15 # Comments: 1
Check a UK supplier at Companies House from Claude Code — and the same tool works for Norway
A new supplier sends an invoice with a company registration number on it. Is the company real, still trading, and up to date on its filings? answers from Companies House by company number — and from Brønnøysundregistrene / Enhetsregisteret (brreg) by organisasjonsnummer (orgnr, org.nr), with the same tool and the same JSON. The prompt Invoice from Tesco PLC, company number 00445790. Is it active, and is it up to date at Companies House? The call Read the s literally. Companies House publishes no VAT status and no employee count, for any company — means "this register does not say", never "no". Norway publishes both: One shape, two registers, honest about which publishes what. The half that's actually useful Companies House publishes each company's own filing due dates, so quotes them...
Before You Grant an AI Agent Access: A Practical OAuth Scope Audit
Meta released Muse, a personal AI agent that asks for access to your email, calendars, payments and health data in order to do things for you — selling a car, booking travel. It competes with agents like OpenClaw and Instinct, and it pairs broad data requests with promises of privacy controls. Whether Muse succeeds depends on whether people decide to trust Meta with that scope. That's a fair question but not the one I care about here. The useful question applies to every agent in this category, including the ones you'll be offered next month: What exactly are you granting, and what keeps working after you stop watching? An agent is not an app This is the distinction most people miss, and it's the whole risk model. An app asks for permission and runs when you open it. Access is bounded by...