2 New Claude Code Features That Matter for L&D Professionals

Claude Code Just Got Two Big Upgrades

If you use Claude Code to build training content, maintain course websites, or generate learning materials, two recent updates are worth your attention. Remote Control (February 2026) lets you continue a desktop coding session from your phone or tablet. Scheduled Tasks (March 2026) lets you put recurring prompts on autopilot so Claude runs them in the background while you do other work.

Neither feature got much coverage in the L&D space, but both solve real problems I hit every week. Here's what they actually do and how I'm using them.

Feature 1: Remote Control

The Problem It Solves

I was knee-deep in a course build at my desk, Claude Code session running in my terminal, files open, MCP servers connected, the whole setup. Then I had to leave for my niece's piano recital. Normally that means closing the laptop, losing context, and spending 20 minutes getting back into flow when I return.

Instead, I typed /remote-control, scanned a QR code with my phone, and kept working from the auditorium while waiting for her turn. I approved a file edit, reviewed the HTML output Claude generated, and sent a follow-up prompt to adjust the quiz logic, all from the Claude app on my phone. When I got back to my desk, the terminal had every change ready and waiting.

What Remote Control Actually Is

Released in late February 2026, Remote Control connects your local Claude Code terminal session to claude.ai/code in a browser or the Claude mobile app on iOS and Android. The key distinction: your session keeps running on your machine. Your files, your project folder, your MCP servers, your environment variables, they all stay local and available. The phone or browser is just a window into that local session.

This isn't a cloud IDE. Nothing moves to someone else's server. Claude reads and writes files on your actual computer while you send prompts from wherever you happen to be. If your laptop sleeps or your Wi-Fi drops, the session reconnects automatically when things come back online.

Starting a session takes one command:

claude remote-control --name "Onboarding Course Build"

That gives you a URL and a QR code. Scan it, and you're connected. You can also type /rc from inside an existing session to upgrade it to remote on the fly.

Why This Matters for L&D Professionals

If you're an instructional designer who also touches code, whether that's building custom web-based training, maintaining a portfolio site, scripting Storyline interactions, or generating course content with AI, Remote Control solves a specific and annoying problem: context switching when you leave your desk.

Here are the scenarios where I've already found it useful:

1. Approving Changes During SME Meetings

I was in a subject matter expert call on my phone when the SME mentioned a correction to the compliance module. I had Claude Code running on my desktop with the project loaded. From my phone, I told Claude to update the section, reviewed the diff, and approved it, all while still on the call. No "I'll make that change when I get back to my desk" delay. The SME saw the update on the staging site before we hung up.

2. Monitoring Long Builds and Deploys

When I publish a new interactive training module, the build process, Vite compilation, asset optimization, Netlify deploy, takes a few minutes. With Remote Control, I can kick off the build, walk away from my desk, and get the completion status on my phone. If the deploy fails, I can read the error and fix it without walking back to my computer.

3. Quick Content Iterations Between Meetings

L&D work involves a lot of meetings. Back-to-back stakeholder reviews, design walkthroughs, pilot feedback sessions. The gaps between those meetings, the 10 minutes waiting for someone to join, the lunch break, the commute, those used to be dead time. Now I use them to send Claude a prompt: "Add alt text to all images in the safety module" or "Generate five more quiz questions for the data privacy section." The work happens on my desktop. I just approve it from my phone.

4. After-Hours Fixes Without Opening a Laptop

A stakeholder emails at 8 PM saying there's a typo on the training landing page. Before Remote Control, that either waits until tomorrow or I open my laptop and get sucked into a 45-minute work session. Now I start a quick remote session from my phone, tell Claude to find and fix the typo, approve the change, and deploy. Five minutes, done, laptop stays closed.

The Security Model Actually Makes Sense

One thing that matters when you're working with training content, especially anything involving proprietary processes, compliance material, or employee data: nothing leaves your machine except through Anthropic's standard encrypted API connection. Your local session makes outbound HTTPS requests only. No inbound ports open on your computer. No files uploaded to a third-party cloud environment.

This is different from Claude Code on the web, which runs in Anthropic's cloud infrastructure. Remote Control keeps everything local. For L&D teams working under NDAs or with sensitive organizational content, that distinction matters.

What It Can't Do (Yet)

Remote Control isn't magic, and it has real limitations you should know about:

Your terminal must stay open. If you close the terminal window or shut down your computer, the session ends. This isn't a persistent cloud agent. It's a bridge to your running local session. Sleep mode is fine though, it reconnects when your machine wakes up.

One remote session per terminal instance. You can't have five people connecting to the same session. It's one-to-one: your phone to your desktop.

Standard permission rules apply. Remote Control doesn't change Claude Code's permission model. By default, you'll still approve actions like file edits and shell commands from whatever device you're connected on. If you've configured auto-accept or custom permission settings locally, those carry over to the remote session too.

Network timeouts. If your desktop loses internet for more than about 10 minutes, the session ends and you'll need to start a new one.

How It Compares to Just Using Claude on the Web

Anthropic also offers Claude Code on the web at claude.ai/code, which runs entirely in their cloud. That's great for quick tasks or working on repos you don't have cloned locally. But for active L&D projects where you have local files, custom MCP servers, specific build tools, and a configured environment, Remote Control is the better fit because it uses your actual setup.

Think of it this way: Claude Code on the web is a fresh workstation in the cloud. Remote Control is your actual desk, accessible from your pocket.

Feature 2: Scheduled Tasks

The Problem It Solves

Every morning I used to open Claude Code and type the same thing: "Check the training site for broken links and missing alt text." Every time I deployed a new module, I'd manually ask Claude to verify the build output. Every Friday I'd run an accessibility audit. Same prompts, same workflow, every single time.

Scheduled Tasks eliminates that repetition. You tell Claude to run a prompt on a schedule, and it just does it in the background while your session is open.

What Scheduled Tasks Actually Is

Released on March 7, 2026, the /loop command lets you schedule recurring prompts that fire automatically at intervals you set. Type this:

/loop 2h check the training site for broken links and report any issues

And every two hours, Claude runs that prompt, checks your site, and reports back. You can also set one-time reminders in natural language:

remind me at 3pm to push the updated compliance module to staging

Under the hood, it uses standard cron expressions, so you can get precise if you want. 0 9 * * 1-5 means "every weekday at 9 AM." But for most use cases, the simple interval syntax — 5m, 1h, 2d — is all you need. All times use your local timezone, not UTC.

How I'm Using It for L&D Work

1. Post-Deploy Monitoring

After I push a new training module live, I set up a quick loop:

/loop 10m check if the Netlify deploy finished and tell me the status

Instead of refreshing a deploy dashboard, Claude tells me when it's done and flags any errors. The loop auto-fires between my turns, so I can keep working on other things in the same session.

2. Automated Content QA

I run a loop that checks my training pages for common issues — broken links, missing alt text, console errors. Setting this up once per session means I catch problems the moment they appear, not three days later when a stakeholder reports them.

3. Build Babysitting

Some of my interactive training modules have long build processes. I'll kick off a build and immediately set a loop to check whether it's finished. When it completes, Claude reports the result and the loop stops being useful, so I cancel it with a simple "cancel the build check job."

4. End-of-Day Reminders

This one is simple but effective. I'll tell Claude: "at 4:30pm, summarize what files we changed today and what's left to do." It fires once, gives me a clean summary of the session's work, and I have my status update ready for the stakeholder email.

What You Need to Know About Limitations

Session-scoped only. Scheduled tasks only exist while your Claude Code session is running. Close the terminal, and they're gone. This isn't a persistent cron service — it's session automation. For durable scheduling that survives restarts, Anthropic points to GitHub Actions or Desktop scheduled tasks.

Three-day expiry. Recurring tasks automatically delete themselves after three days. This prevents forgotten loops from running forever, but it means you'll need to recreate long-running schedules.

No catch-up for missed fires. If Claude is busy with a long response when a scheduled prompt comes due, it fires once when Claude finishes — not once per missed interval. So timing isn't precise to the second.

50 task limit. You can have up to 50 scheduled tasks per session, which is more than enough for any workflow I can imagine.

Both Features Together

Here's where it gets interesting. Remote Control and Scheduled Tasks work together. I can start a Claude Code session at my desk, set up a few scheduled loops for monitoring and QA, then walk away and check results from my phone via Remote Control. Claude keeps running the scheduled prompts on my desktop while I review the output from wherever I am.

That combination — automated background work plus mobile access — turns Claude Code from a tool I sit in front of into a tool that works alongside me throughout the day.

Getting Started