Grégoire Clermont

439 posts

Grégoire Clermont

Grégoire Clermont

@gregclermont

building security at https://t.co/mBie3u14lP

France เข้าร่วม Nisan 2010
1.1K กำลังติดตาม559 ผู้ติดตาม
Jessie Frazelle
Jessie Frazelle@jessfraz·
while we are at it, anyone at microsoft gov cloud entra, i submitted a bug there too, its a bit more complex but having someone who gives a shit would be nice
English
5
0
56
7.1K
Jessie Frazelle
Jessie Frazelle@jessfraz·
who do i know at github security, we found a bug and i want to make sure it goes to someone who gives a shit and not just an ai bot
English
16
8
286
52.3K
Grégoire Clermont รีทวีตแล้ว
Neil Renic
Neil Renic@NC_Renic·
By using AI for writing, you’re robbing yourself of the authentic writer’s experience of not writing
English
99
8.4K
57K
827.6K
Grégoire Clermont รีทวีตแล้ว
Squiblydoo
Squiblydoo@SquiblydooBlog·
From 2020-2024, I tracked the SolarMarker malware, and in 2024, monitored a self-infection for months to learn their actions-on-objectives: on-device fraud. I didn't publish the details of my months long investigation until now. Check the link the the attached comment.
Squiblydoo tweet media
English
2
15
65
8.4K
Grégoire Clermont รีทวีตแล้ว
Squiblydoo
Squiblydoo@SquiblydooBlog·
CertCentral is now TheCertGraveyard[.]org & CertGraveyard[.]org. The CertCentral API returns an error directing to use the new domains. Please give me a like or a share to get the word out. Also use the site to report and investigate certificates used to sign malware. :)
Squiblydoo@SquiblydooBlog

I'm being required to give up the domain CertCentral[.]org; and the change has to happen by Monday. I'm noodling on alternative names. Keep an eye out for the change.

English
5
47
82
19.6K
Grégoire Clermont
Grégoire Clermont@gregclermont·
@IAMERICAbooted One vector that I'm still worried about is OAuth phishing with first-party Microsoft app (e.g. volexity.com/blog/2025/04/2…). As almost all the CA checks are performed on the victim's side, I've found this technique hard to mitigate. Do you have recommendations?
English
0
0
2
1.2K
EZ
EZ@IAMERICAbooted·
Do you know why I focused most of time on entra conditional access when I was consulting? Because Jef Kazimer at Microsoft constantly reminded me what you can do with CAPs and how they mitigate almost all M365 initial access. And he was right. However, if you are allowing users to access M365 as BYOD (not phones, we are not talking about MAM-WE here) and you are not forcing authentication contexts requiring phishing resistant authentication that mitigates downgrades (hardly anyone), your CAPs alone will not save you from uploads, and neither will SharePoint settings. There is the forgotten beast called Power Platform. In consulting, I have yet to see any orgs that dont have API gaps due to resource targeting in CAPs. This is why you need exclude by exception and when you do exclude, you need to understand that exclusion is creating an API gap.
English
8
12
116
9.2K
Grégoire Clermont
Grégoire Clermont@gregclermont·
@ItsReallyNick Any chance Microsoft might officially document UserAuthenticationMethod at some point? I realize it's not your remit, but if this ever comes up in conversation with the right folks internally, there's definitely interest in the community!
English
0
0
0
50
Grégoire Clermont
Grégoire Clermont@gregclermont·
@ItsReallyNick Thanks, Nick! Your tweets and GitHub comments were the only information I could find coming from a Microsoft employee (or anyone, really) about this field! That definitely gave me some confidence that this bitfield interpretation was credible.
English
1
0
1
54
Grégoire Clermont รีทวีตแล้ว
Sekoia.io
Sekoia.io@sekoia_io·
Our latest technical deep-dive unravels the mystery behind the opaque numeric codes (16, 272, 33554432, etc.) you see in #Microsoft365 audit logs. buff.ly/w04flj1
Sekoia.io tweet media
English
1
7
15
1.7K
Grégoire Clermont
Grégoire Clermont@gregclermont·
@ZackKorman I have a pet theory that the reason why Kusto got so versatile and powerful is because the Entra/Azure/M365 logs are so convoluted and impractical. Instead of fixing the logs, they built a crazy query language.
English
1
0
1
50
Zack Korman
Zack Korman@ZackKorman·
I can assure you, Microsoft has never once thought “how do we reduce noise and improve the usability of the audit log”
English
5
1
70
1.2K
Zack Korman
Zack Korman@ZackKorman·
Consensus view on Reddit is that Microsoft probably decided to not log Copilot events to “reduce noise” and make the audit log “more useful”. And that’s how you know none of them have ever seen an audit log before.
English
12
24
363
17.3K
Grégoire Clermont รีทวีตแล้ว
Justin Elze
Justin Elze@HackingLZ·
I love Cloudflare, but the number of phishing sites I find sitting behind Cloudflare or utilizing Turnstile is excessive.
English
24
11
346
28.2K
Grégoire Clermont รีทวีตแล้ว
Michael 🆘
Michael 🆘@matonis·
Hey everyone, I’m proud to announce that I'm hosting a new conference called State of Statecraft (🆘) October 28 in Brussels, Belgium. SOS exclusively discusses state-sponsored operations. Speakers! Our CFP closes in 2 weeks!!! -> Apply here: stateofstatecraft.com/cfp
Michael 🆘 tweet media
English
2
21
64
9.8K
Grégoire Clermont รีทวีตแล้ว
L0Psec
L0Psec@L0Psec·
New RE Blog Post: RustyPages-Pt1 the-sequence.com/rustypages-mal… We RE a Rust dropper, that sets persistence and runs the downloaded next stage, queries @patrickwardle's tools, and quiets notifications. We included relevant IOCs as we continue our analysis of the loader for Part 2. :)
English
1
34
104
13.1K
Grégoire Clermont รีทวีตแล้ว
Fabian Bader
Fabian Bader@fabian_bader·
New Microsoft Graph based API for response actions in #MDI Disable, Enable, ForcePasswordReset and RevokeAllSessions finally available for your automations. #new-graph-based-api-for-response-actions-preview" target="_blank" rel="nofollow noopener">learn.microsoft.com/en-us/defender…
Fabian Bader tweet media
English
2
25
145
8.4K
Mehmet Ergene
Mehmet Ergene@Cyb3rMonk·
⚠️ Threat hunting and detection engineering using time series anomaly detection methods is so powerful but quite tricky. Here is why: - Users/endpoints don't have a reliable baseline - First seen activities don't fit time series anomaly logic by default (with some modifications, it can) - Compute resources limit the historical data that can be analyzed, which directly impacts what the logic outputs with default parameters. That's why I have multiple lessons to explain how to develop robust detection logic. #ThreatHunting #DetectionEngineering
Mehmet Ergene tweet media
English
2
12
78
6.8K
Grégoire Clermont รีทวีตแล้ว
Sekoia.io
Sekoia.io@sekoia_io·
A few weeks ago, we published our global analysis of Adversary-in-the-Middle #phishing threats, providing actionable intelligence on multiple #AitM phishing kits. This report includes 11 sheets covering the most widespread #AitM phishing kits as of Q1 2025.
Sekoia.io tweet media
English
1
22
53
6.7K