FOSS Weekly #26.09: Linux Mint Shortcuts, OpenClaw Alternatives, Ladybird’s Rust Move, Super Productivity and More

Issue 26.09 of the FOSS Weekly digest landed at the end of February 2026, and it is the kind of issue that does not have a single lead story. There is Linux Mint keyboard shortcuts, a list of OpenClaw alternatives that run on single-board computers without a Raspberry Pi, a look at Ladybird’s move into Rust, a productivity app called Super Productivity, and a pair of AI stories that are worth your time even if you are tired of hearing about AI. Here is what stood out to me, and what I am going to keep an eye on.

The Linux Mint shortcuts piece

The Linux Mint shortcuts guide is a perennial FOSS Weekly feature because every six months a new batch of users installs Mint and goes looking for the keybindings that make the desktop feel fast. The 2026 edition covers the same ground as earlier ones, but it adds a few XFCE-specific shortcuts that I had been missing for years.

A few that I lifted directly into my muscle memory this week.

• Super key alone opens the Whisker menu (Mint’s app launcher, similar to the Windows Start menu or macOS Spotlight). I had been mousing into it for years.
• Super + D shows the desktop. Standard, but Mint’s behavior is to actually minimize all windows, not just show the desktop, which is the right call.
• Super + E opens Thunar (Mint’s default file manager). One keystroke, no dock clicking.
• Ctrl + Alt + T opens a terminal. The Linux classic. If you are on Mint and you do not use this, you are missing the easiest win in the whole desktop.

None of this is breaking new ground, and the FOSS Weekly team would be the first to admit it. The value is in the reminder. Most of us have muscle memory from our first Linux install, and it lags the desktop by years. A 10-minute pass through the shortcut list is the kind of low-effort, high-return cleanup that pays off the rest of the month.

The OpenClaw alternatives for SBCs

The OpenClaw alternatives list caught my eye because I have been running OpenClaw on a Raspberry Pi 5 for about a year, and I have been quietly looking for the exit. OpenClaw is a good fit for a Pi 5 with 8 GB of RAM, but the rest of my fleet is a mix of Orange Pi 5 Plus boards and a couple of Radxa Rock 5B units, and OpenClaw does not run cleanly on any of them. The driver story for the Rockchip SoCs (systems on chip, the integrated processor packages used in many SBCs) is rough, and the OpenClaw maintainer has been clear that the Pi 5 is the reference platform.

Options on the list are mostly lighter-weight alternatives that do not need the full Pi 5 memory budget. I have not run them all, but two of them are now on my shortlist.

• Option A. A small Go-based agent that does the same job as OpenClaw for the specific subset of tasks I actually use, with a memory footprint under 100 MB. The trade-off is that it is much less general-purpose, and you cannot bolt on the plugin set.
• Option B. A web-first agent that runs in a browser tab and talks to local services over a WebSocket. The “local” part is real, the data does not leave the device, and the memory footprint is the browser’s, not the SBC’s.

Both are worse than OpenClaw for the general case, and both are better than OpenClaw for my specific fleet. That is the right shape of trade-off to be making in 2026.

Ladybird and the Rust move

The Ladybird browser project announced in February 2026 that it is moving parts of its implementation from C++ to Rust. The reaction in the FOSS Weekly comments was the usual mix of “this is a good idea” and “this is going to be a five-year migration.” Both are correct, and the digest does a good job of being clear about which parts are moving first.

First up are the parser and the layout engine, which are the two places where memory safety bugs hurt the most. The rendering pipeline and the JS engine stay in C++ for now, because rewriting those is a different scale of project. Andreas Kling has been clear that the Rust port is a layered migration, not a flag-day rewrite. Each component lands when it is ready, and the older C++ stays in place until the new Rust code is feature-complete.

I am cautiously optimistic. Ladybird is a fun project to watch because Kling is shipping a real browser from scratch, and any time a project like that takes memory safety seriously, the rest of the community benefits. The risk is that the migration slows down the release cadence at exactly the moment when Ladybird needed to keep moving. We will know by the end of 2026.

Super Productivity

The Super Productivity app is a self-hosted task manager that integrates with Jira, GitHub, and a Pomodoro timer (a time-boxing technique that breaks work into 25-minute focused intervals separated by short breaks). I had heard of it but never installed it, and the FOSS Weekly walkthrough was the push I needed.

Setup is straightforward. The app runs as an Electron (a framework that lets web apps run as native desktop applications) binary on Linux, macOS, and Windows, or as a Docker container if you want it server-side. The task list is plain Markdown, the timer is built in, and the integrations with GitHub and Jira are real, not toy stubs. I have been using it for a week and the thing I appreciate most is that it does not phone home. The app does not send telemetry (automated diagnostic or usage data sent back to the developer), the GitHub token is stored locally, and the Docker container does not open any ports you did not configure.

A few things to know if you install it.

• The Pomodoro integration is the killer feature. It is the only task manager I have used where the timer actually works the way I want it to, including the long-break scheduling.
• The Jira integration is good but opinionated. If your team uses Jira in a way that does not match Super Productivity’s mental model, you will fight it. Mine does, and it works.
• The Docker image is the right way to run it server-side. The desktop binary is fine, but the Docker setup is what you want for a team.

It is not the prettiest app in the world, and the UI is functional in a way that some people will love and some people will not. For a self-hosted, telemetry-free task manager that integrates with the tools you already use, it is the best option I have found in 2026.

The AI stories that are not about coding

The two AI stories in the digest are not about writing code, and that is why I am flagging them. The first is about Anthropic’s claim that its AI can modernize legacy COBOL codebases. The second is about a research paper showing that AI models can de-anonymize your online identity from a small number of posts.

On the COBOL side, the revenue impact is not abstract. IBM’s stock dropped 16% in a single day on the news, because IBM still makes real money maintaining mainframe systems that run banks, airlines, and critical financial infrastructure. If Anthropic’s claim holds up, the modernization market is a lot larger than people thought, and a lot of the maintenance contracts that have looked unassailable for forty years are not. The work to validate the claim is the next twelve to eighteen months, and that is when the market is going to react.

The de-anonymization piece is the one I think most people will skip over and regret. A research paper showed that off-the-shelf AI models can take a handful of anonymous online posts and figure out who wrote them. The cost is a couple of dollars per target. This is not a theoretical attack. It is a tool that exists today, and the barrier to using it is essentially zero.

If you write online under a pen name, or you post in communities that you would rather not be doxxed (having your real identity publicly revealed without consent) from, the safe assumption is that this attack is available to anyone who wants to use it. Pseudonymity on the open internet is a much weaker property in 2026 than it was in 2024, and the tools that break it are getting cheaper every quarter.

Trade-offs

A weekly digest is by definition a tasting menu, and the trade-off is that nothing gets a full deep-dive. The Linux Mint shortcut list is a reminder, not a tutorial. The OpenClaw alternatives are a starting point, not a full evaluation. The Ladybird Rust move is an announcement, not a roadmap. Reading a digest like this is useful for staying oriented, and it is no substitute for actually installing the tools and trying them.

There is also a meta trade-off in the AI stories. The COBOL modernization story is bullish for the AI industry and bearish for the people who do COBOL maintenance for a living. The de-anonymization story is bullish for law enforcement and bearish for everyone who has ever posted under a pseudonym. Both are real, and both are happening at the same time, and the digest does not pretend they are the same kind of story.

What I would tell past me

If I could send a message to the version of me that skipped FOSS Weekly digests for a year, I would say four things.

• Read the roundup, then pick one thing to install. The whole point of a digest is to surface the things you would not have searched for. Install one of them each week, and you will end the year with twelve new tools instead of zero.
• The keyboard shortcut pieces are not beneath you. I have been using Linux for years and I still learn a new shortcut every time. The 10-minute pass pays off forever.
• Take the AI de-anonymization story seriously. It is the kind of thing that is true today and that nobody will tell you about in a year. Adjust your online behavior now, not after it matters.
• OpenClaw on a Pi 5 is fine. OpenClaw on a Rockchip SBC is not. Pick the platform that matches the tool, or pick the tool that matches the platform. Do not fight the driver stack.

Issue 26.09 is a good snapshot of where the FOSS community is in late February 2026. The throughline is that the work is moving, the maintenance burden is real, and the tools that exist today are better than the tools that existed last year, in the specific ways you would want.