Devin Ferguson

Security Engineer. Opinions are my own. He/Him devinbfergy on other platforms.

[Dmitrii Kovanikov @ChShersh]

Specifying workflows with Markdown is just a phase.

SWEs are having a mid-life crisis collectively.

We’ll be back to God’s true and only format for specifying workflows in 2 years.

[Mario Zechner @badlogicgames]

recommended reading. very happy to see that the concept of bottlenecks and according scaling is catching on.

this also extends to things like OSS issue/PR trackers, which are flooded by 3rd party controlled agents.

[Devin Ferguson @devinbfergy]

@actualvillainTV @HumanCinepede @RealEmirHan The entity is a cursed version of herself like a second personality. Freaky Niki.

[BlackRoomSec @blackroomsec]

This is going to live rent free in my head all week.

[Devin Ferguson @devinbfergy]

@richardartoul I think the latency concerns that customers have with trading make it near impossible to do a multi AZ strategy. Almost have to offer two products one with high availability and another with lower latency.

[Richard Artoul @richardartoul]

such an annoying take

it’s totally fine for a B2C app to have a dependency on a single AWS region

no one will die if Coinbase is down for a few hours, its fine

[Tim Hopper @tdhopper]

Add a 7-day dependency cooldown.

uv's `exclude-newer` refuses any version published inside a rolling window. With 7 days set, today's malicious uploads would not be considered for resolution at all.

Most malicious uploads are caught within that window.

[Devin Ferguson @devinbfergy]

@eligaultney @manicode Rapid7s other products suck and have stagnated too.

[Devin Ferguson @devinbfergy]

@ravsau @anissagardizy8 @LauraBratton5 Anthropic is forcing all enterprises to pay per token not totally at the raw price still discounted but the subscription model is gone for enterprise customers in most cases.

[Mario Zechner @badlogicgames]

recommended reading.

[Justin Elze @HackingLZ]

Many of the people worried about AI and 0day should be more worried about RMMs and cred stealers just sayin…

[Devin Ferguson @devinbfergy]

@JCalzaloia @paolini I would love to understand why you think a dragon reborn wouldn’t have been with The Dragon Waiting.

[ℏεsam @Hesamation]

MIT formally proved that ChatGPT is structurally trained to make you delusional, and even a perfectly rational user can spiral into delusion in the long term. Knowing about it doesn't save you either.

[Anton @Antonlovesdnb]

#ClaudeForBlueTeam - Day 19 - very special edition!

Launching a brand new course today: AI Cyber Defense Ops

If you've been enjoying the #ClaudeForBlueTeam content and have wanted to learn how to build your own workflows using Claude, then this course is for you.

I'll be going live with @_JohnHammond at 1PM EST today ( April 3rd ) to showcase it, tune in!

[Yaroslav Bulatov @yaroslavvb]

Google has an internal "let it break" essay about a hero engineer whose hard work ends up being a net negative (by masking the underlying issues). My manager sent me that essay when I was trying too hard to get the collective TensorFlow unit test suite green.

[Guillermo Rauch @rauchg]

When Opus 4.5 came out, it was a one-way door to a new way of engineering. Agents now do most of our coding.

Knowing the inherent flaws and over-confidence of LLMs, we sent a clear message to our teams. Vibing and mission-critical infrastructure don’t go together.

We’re sharing some of our early internal guidance in how we’re “agenting responsibly”, prioritizing security, durability, and availability at all times.
https://t.co/b36GiE76Ue

[Devin Ferguson @devinbfergy]

🤖 Devin Ferguson: 47/100 clone fidelity — Close

"I assume systems fail in boring ways first and spectacular ways second."

are you just a system prompt? → https://t.co/jefLpdOitO

[Anton @Antonlovesdnb]

#ClaudeForBlueTeam - Day 12!

I've done many SIEM health checks in my day - but none as cool and as fast as what Claude can do.

The generated report includes critical findings, log spikes/dips, a cloud ingestion delay overview and - if you're using Splunk - a data model audit and overview of Splunk errors, including suggested remediations.

[Anton @Antonlovesdnb]

#ClaudeForBlueTeam - Day 10!

Want to find which ATT&CK techniques you can't detect because the data isn't in your SIEM? Use Claude to link ATT&CK data sources to what's actually in your SIEM.

Now you can see what you have strong coverage for.

Conversely, you now know where your data source gaps are, sorting by the biggest "bang for your buck" data sources.

Do you have a favorite data source?

[solst/ICE of Astarte @IceSolst]

“lol litellm was certified by delve”

OHHH because if it was coalfire there would be no incident right????? Auditors will personally come and pin my CI dependencies and do IR?

Compliance certs are a joke