fractorial

From TFA:

> Note that disabling MPTCP on a program-by-program basis is a fragile mitigation. Accidentally leaving a single MPTCP listener exposed will still leave a system vulnerable. Further, it assumes that all viable exploits for this bug depend on MPTCP. Again, we strongly recommend upgrading to a patched kernel version

First time?

s/saying/implying/

As someone who rolled and daily drives a harness leaner than Pi, there’s still a use case for both.

Similarly to “one language is best,” saying “one harness is best” is shortsighted in the best possible case.

That being said, Claude Code is bloatware.

Seeing Atlanna inspire an open weight model was not on my 2026 bingo card.

Seriously hope you get that trace because as much as I like to dunk on Claude Code as a harness, it’s definitely not as bad as OP stated, ha.

Would be cool to see a not slop version of this.

> Downstream runtimes that ship their own trust stores (Go, Node.js, Electron, Python certifi, curl, etc.)

At least two of these don’t ship a trust store.

Works every time 97% of the time.

Ah, yes that seems to be what URGENT track self-documents as.

What about it makes it so interesting?

If you had watched Grant for years, like I have, you would surely see how this is misinformation at best.

Curious: What are the trivial agent vs. human detection mechanisms?

How is Anthropic different? You still cannot use a subscription in a custom harness in the same way Codex allows you to.

And, no, wrapping claude -p or any other “allowed” use wherein you don’t control the agent loop isn’t the same.