Rami

🚨 We scanned GitHub and found *hundreds* of valid secrets, 4 of the top 5 were AI-related:
HuggingFace, Azure OpenAI, Weights & Biases, and Groq.

Read more:
www.wiz.io/blog/leaking...

Securing open-source credentials at scale | Google Cloud Blog We’ve developed a powerful tool to scan open-source package and image files by default for leaked Google Cloud credentials. Here’s how to use it.

> We've set up a web endpoint so vetted ... security researchers can submit suspected exposed credentials for review
> To report exposed Google Cloud credentials, please contact gcp-credentials-reports@google.com

cloud.google.com/blog/product... really buried the lede!

Hardening GitHub Actions: Lessons from Recent Attacks | Wiz Blog Build resilient GitHub Actions workflows with insights from real attacks, missteps to avoid, and security tips GitHub’s docs don’t fully cover.

In light of recent GitHub Actions incidents (Ultralytics, tj-actions...), I wrote up a practical guide to hardening for @wizsecurity.bsky.social

Covers permissions, secrets, 3rd-party Actions, ++

Use it to avoid learning these lessons the hard way:
www.wiz.io/blog/github-...

MCP and LLM Security Research Briefing | Wiz Blog Explore the evolving Model Context Protocol (MCP), its security risks, and how to prepare for safe adoption as LLMs connect to external systems.

Synthesized 20+ sources and internal @wizsecurity.bsky.social expertise to come out with a comprehensive guide to MCP security

Today's options, and tomorrow's possibilities

www.wiz.io/blog/mcp-sec...

Quadruple Supply Chain Attack, IngressNightmare Exploited, and Rumors Abound Podcast Episode · Crying Out Cloud · 04/09/2025 · 29m

🎙️ New episode!
Our own @ramimac.me helps dive into GitHub supply chain attacks, IngressNightmare, and Oracle breach rumors.

Tune in for the latest cloud security insights!
🎧 podcasts.apple.com/us/podcast/q...

CloudTrail Network Activity Events for AWS VPC Endpoints | Wiz Blog How AWS VPC Endpoint CloudTrail logs can help you troubleshoot endpoint policies and strengthen your network's security against data exfiltration.

It's been awesome getting to team up with @ramimac.me to dig into a new AWS feature! Read our thoughts on AWS's new CloudTrail network activity events (aka VPC endpoint logs): www.wiz.io/blog/aws-vpc...

GitHub Action supply chain attack: reviewdog/action-setup | Wiz Blog A supply chain attack on tj-actions/changed-files leaked secrets. Wiz Research found another attack on reviewdog/actions-setup, possibly causing the compromise.

Turns out when you investigate a compromised Github Actions you ... find another compromised Github Action:

www.wiz.io/blog/new-git...

Very fun to help put final polish on this report in week 3 at Wiz - anecdata is fun, data is funner :)

🧙 Why I’m Joining Wiz I’m joining the leading cloud security startup, hoping to “work for the Security Industry, at Wiz.”

New year, new job!

I've joined the amazing @wiz_io research team

My goal is the "work for the security industry, at Wiz"

I wrote a blog post explaining why, and what that means:
ramimac.me/joining-wiz

How to Say “No” Well Security’s pivot from ‘Department of No’ to ‘Department of Yes’ misses the real lesson - how to say ‘No’ the right way.

Lately, every BSides seems to have a talk on reframing security teams as a “Department of Yes”

We don’t hear nearly as much about the value of a well-considered, strategically deployed “No”

I've pulled together guidance on giving a better, more constructive No:
ramimac.me/saying-no

Keep an eye out for notices - AWS RDS Protection for Guardduty seems to have had some issues collecting logs.

Unclear how pervasive this was!

Poisoning the SSM Command Document Well Responsibly disclosing risks in using SSM Command Docs for software distribution.

reminds me of ramimac.me/poisoning-ss... 😉

State of ‘State of Cloud Security’ Reports: Insights or Self-Owns? Dozens of hours reading State of Cloud Security reports that I think miss the mark.

One recent report highlighted that roughly a third of their customers have “at least one cloud workload that is publicly exposed, critically vulnerable and highly privileged.”

If you’re this vendor, should I really buy your product?

ramimac.me/state-of-clo...

I've spent dozens of hours reading State of Cloud Security reports

You know, the ones that use data from their CSPM product

And I've realized the findings substantially reflect how well that tool helps customers secure their clouds

I wrote up some examples, both good and bad (🔗 in 🧵)

“Founder Mode” and the Art of Mythmaking I’ve never been good at “hot takes”. Anyone who knows anything about marketing can tell you that the best time to share your opinion about something is when everyone is all worked up about it. Hot …

I (finally) wrote up my thoughts on "Founder Mode" and the Brian Chesky morality tale about how he turned around Airbnb company culture.

This has made it into the Silicon Valley water table; it must be dealt with. There are some good nuggets within; let's dig them out.

charity.wtf/2024/12/17/f...

www.cybok.org/media/downlo...

New Threat Vector Unlocked

1. Find the Crunchbase page of a cybersecurity company that just raised VC funding
2. Change the page details (which anyone with a Crunchbase account can do) to a personal CashApp page
3. ????
4. Profit! (?)

YouTube video by LimaCharlie A decade of defense: securing the largest US crypto exchange | Philip Martin | MSSN CTRL 2024

Somehow <50 people have caught this talk from Coinbase's CSO??

His core advice:
1. Make lives easier - e.g roll out yubikeys
2. Define Security Invariants
3. Plan & Practice IR
4. Balance Risks & Threats
5. Security is a People Problem - use focus groups for new controls!

youtu.be/BPh4Hc3TH74

Just keep dm'ing them to me so I can bully you into publishing 😈

Interesting research out of AWS!

> IAM-PolicyRefiner, a tool that automatically synthesizes refined AWS IAM access control policies from access logs

> fast (<5s per policy), effective and does not overfit

Not open source, but maybe a sign of things to come?

assets.amazon.science/cf/bc/58e56f...

IIRC, re:Inforce was a bust for security announcements.

I wish some of these announcements got time center stage there, versus being pre:Invent announcements now.

Hopefully re:Invent makes space alongside the GenAI noise!

I've been chatting a lot re:when to make the first security hire" recently

I've come up with a Rule of Thumb:
Hire your first security person when security is an unavoidable distraction from scaling your business

ramimac.me/start-security

h/t @grims.bsky.social & @mag00.bsky.social

This webinar will be more relay-race than sparring match when it's with folks like @nanook.bsky.social and @jamesberthoty.bsky.social!

Don’t click: towards an effective anti-phishing training. A comparative literature review - Human-centric Computing and Information Sciences Email is of critical importance as a communication channel for both business and personal matters. Unfortunately, it is also often exploited for phishing attacks. To defend against such threats, many ...

I think it's pretty easy to dig up research that ties phishing simulations to a decrease in metrics like Click Through Rate or improvement in "awareness" - but I'd argue those aren't the metrics that matter!
Ex. hcis-journal.springeropen.com/articles/10....

I love when new research comes out to back up my "phishing training is bad practice" priors:

www.computer.org/csdl/proceed...

I track the latest evidence against phishing simulations: rami.wiki/phishing-sim...

Platform Engineering Until recently, infrastructure was the backbone of organizations operating software they developed in-house. But now that cloud vendors run the computers, companies can finally bring the benefits of a...

Wondering “What Does Success Look Like?”

Check out the book!

There is incredible depth, detail, and color provided that I can only hint at in this format: www.oreilly.com/library/view...

Thank you to Camille and Ian for sharing their expertise

“Saying “No” Without Ruining the Relationship”

- “Not yet, priority call” → give options and how they can help
- “Not yet, technical call” → take time to explain the details
- “No, product strategy call” → accept shadow IT or alternative platforms
- “No, technical call” → beware magical thinking

high power + low interest = keep satisfied low power + low interest = monitor with minimum effort high power + high interest = manage closely low power + high interest = keep informed

Use the handy-dandy power-interest grid to prioritize stakeholder management!

📦 Tips & Tricks for Migrations

* Minimize Glue and Limit Variation
* Transparent Migrations
* Usage Metadata
* Automate as much as possible
* Focus documentation on on-ramps and off-ramps

“If you don’t end up regretting your early technology decisions, you probably overengineered. -Randy Schoup”

Don't do a v2 (Pioneer mindset, new features, and scale don't mix!), just change the plane's engine in midair.

Rearchitect with a focus on “reliability, features, efficiency, and security”