I don't want to be an alarmist, but we are probably fucked if a single company consumes as much energy as countries like New Zealand or Morocco. Just saying.

🙋‍♂️ want a zero-config backup of your github repos and think @tangled.org is pretty cool? then synchub.to is for you! 👉 it's a github app that automatically syncs your public github repos to tangled ✨ let me know what you think! 🙏

Alright! It is done! Finally made a GitHub app for AgentScan so you can just install it, select your repositories, and more easily keep automation lovers out of your projects. Give it a spin and drop me some feedback on GitHub

Is the world healing or running out of money? agentscan.tools/health

For all the IDE-nerds out there who like a good font and theme and still enjoy looking at their code, I've just found my new favorite font. So good. madmalik.github.io/mononoki/

agentscan.tools now has a breakdown for each flag to better understand classification reasons and spot potential bugs.

I think we are 5 PRs away from burning the entire planet.

This is how we handle clankers at home! I've just released the AgentScan GitHub action with bounty detection and immediately caught one.

So you guys are calling an AI that calls multiple AIs, which then opens a PR with the code that is then reviewed by another AI, which then triggers the previous AI(s) to implement those changes, push them, and repeat until the PR is approved, then merge?

GitHub just introduced PR limits, with one notable exception: draft PRs are excluded from the count.

This is peak 2026! Infinite loop of AI congratulations on AI contribution. LGTM!

In case you don't remember who you are

We could always choose quantity over quality, speed over care, competition over collaboration. There is no new paradigm. The other side just got a thousand times louder. And our tools are not yet ready to surface back the folks who keep working together to build a better future for all of us.

We've been running daily scans across 17 repositories to understand how automations are shaping the GitHub ecosystem: tracking which PRs are AI-driven and how effectively teams are closing them. I cannot thank @graphieros.npmx.social enough for all his contributions.

Oh man! We were almost making 3 full weeks!

Finally took the time to donate to two OSS projects I genuinely rely on: @npmx.dev and @e18e.dev. If you've been meaning to do the same for a project you love, there's never a bad time to give back.

If you're feeling burned out because your motivation is fading due to today's tech news or your job, join communities like @npmx.dev, @e18e.dev, and other OSS projects to get your drive back. Stay away from AI agents and interact with real people. It's one of the best things that happened to me.

Yes, you can still be a human and have normal conversations with people.

🙋‍♂️ do I know any oss friends who are facing unwelcome LLM-generated PRs or comments? .... want to fight back?

For all those entitled open source code users, here's how to use the Fork button.

Winter is coming.

"It's very bloaty, and there's a lot of copy-paste, and there's awkward abstractions that are brittle, and it works, but it's just really gross" - Andrej Karpathy I couldn't have summarized the past five years any better.

Just released AgentScan action v1.6.0, now including support for the workflow on issues as well. Thank you @nodeland.dev github.com/marketplace/...

Today marks about 3 months since patient zero, Kai Gritun, was publicly spotted spamming OSS projects using AI. What a time to be alive. Happy Clankers' Day!

Just added support for external lists of flagged accounts on agentscan.tools Thanks @heyandras.dev for this awesome list!

Whenever I get a question that begins with "Did GitHub just break ..." or "Did NPM accidentally ...", my usual response these days is yes unless told otherwise.

We made a fake repo with fake bounties, and the bots are applying fake PRs, so we know who is fake, and we can ban them from the Coolify repo. IQ over 1000

After a few days of data collection, a new page on AgentScan was just published, featuring a daily snapshot of the overall ecosystem health. Your daily dose of clankers. Big thanks to that mythological creature that is @graphieros.npmx.social ❤️🙏 agentscan.tools/health

It should be mandatory to approve a PR using a GIF.

Big thanks to the one and only @graphieros.npmx.social for these super cool graphs showing accounts activity bursts on agentscan.netlify.app

In the era of basically weekly supply-chain attacks on npm packages, lots of developers not even know what the AI is installing for them. That freaks me out more than the supply-chain attack itself.

AgentScan's community-maintained automations/spammers list has an RSS feed available if you find it useful. agentscan.netlify.app/feed.xml

You can now have inline account analysis directly on GitHub. The extension update is pending approval, which may take time. In the meantime, to use it and provide feedback, clone the repository and install the extension manually 🙏 github.com/MatteoGabrie...

I think I cracked it. There's a clear link between how much slop you create in OSS projects and the length of your GitHub profile's README page. It's quite a striking correlation, and I'm not even joking.

I'm working on a little something something. I might be able to get the account analysis directly in issues and PRs, so it's easier to quickly glance at automated or flagged accounts.

We are in a dark period where PRs from real contributors might get auto-closed because maintainers no longer trust contributions to be genuine. Go play some COD if you don't care about the community, and leave open source to people who actually love it.

We might have hit peak-2026 with just one PR.

[developer submits a slop PR] me: "You’re in a desert, walking across the sand, you see a tortoise crawling toward you. You flip it onto its back, its belly exposed in the hot sun, legs struggling to turn over, but it can't without help. But you’re not helping."

Happy that the multi-select feature is finally live on @npmx.dev along with lots of new features, bug fixes, and other updates. github.com/npmx-dev/npm...

Alright, the first version of the GitHub action is published. If you'd like to try it out, please share some feedback. Hopefully, this will help a little with guarding PRs in these dark days of automation craziness. github.com/marketplace/...

If you're looking for some inspiration for your website, look no further. I'm having a blast going around deadsimplesites.com

Just added a `detect-agent` workflow to Clack to tag fully automated PRs. Based on @matteogabriele.bsky.social's agentscan.netlify.app project which surfaces suspicious patterns from account activity. No auto-close for now, but it's an important signal to have. github.com/bombshell-de...

we're all here 🤍

Ok, the Chrome extension has been approved and is open source. It's a first iteration, and I hope to make it more insightful in the future. For now, a link will be displayed on the account page so you can go directly to AgentScan's profile page. 🤖🚫 chromewebstore.google.com/detail/npnlp...

One key takeaway from the recent Vue conference is the importance of human interactions, especially in open source. Contribute to open source, value human connections, have fun, and please leave automations at the door.

Hear me out: what if we go back to making cool apps with great features, and we don't bother telling each other whether they use AI? What if nobody needs to care what powers what?

AI agents flooding PRs should be opt-in, not something users struggle to disable.

Just submitted this small Chrome Extension to the store!

Fun fact: I based the scoring system on how to determine whether GitHub events are likely automated, using @danielroe.dev contribution patterns. If you code more than him, you must be an AI agent.