DML (@damipedia) header
DML (@damipedia) profile picture

MUFC | Cybersecurity | AppSec Engineer | The quicker you let go of old cheese, the sooner you find new cheese.~~~~~ Warrior and Survivor

Mars
Joined May 2011
Photo by DML (@damipedia)Photo by DML (@damipedia)Photo by DML (@damipedia)Photo by DML (@damipedia)Photo by DML (@damipedia)Photo by DML (@damipedia)Photo by DML (@damipedia)Photo by DML (@damipedia)Photo by DML (@damipedia)Photo by DML (@damipedia)
An update on Northflank's October 30 security incident. We’ve been working with independent forensic firm FTI Consulting to investigate. We’ve now published our full report, covering what happened and the changes we’re making. The key findings: 1/ On October 30, an attacker exploited a Northflank API vulnerability and exfiltrated secrets affecting 1.5% of customer teams, representing 1.8% of stored secret values. 2/ The investigation has found no evidence of access to customer workloads or workload data. 3/ We fixed the vulnerability within 82 minutes of receiving the first alert. By the time we posted the status update on October 30, the attacker was no longer in the environment and it was safe to rotate credentials. 4/ We’ve strengthened detection and automated blocking, and are continuing platform hardening and AI-assisted red-teaming. I’m deeply sorry for the worry, disruption and extra work this has caused. Thank you for the patience and kindness you’ve shown our team. We know rebuilding your trust takes sustained work, and we’ll keep sharing our progress. A huge thank you to the Wiz incident response team and Ami Luttwak for their support and expertise throughout this period. northflank.com/blog/incident…
5
8
3
34
4,383
CLOUD was never this easy to understand‼️
2
26
268
12,444
There’s no shortage of great books on security controls or even the basics of running a security program. But, I always wanted a book on how to lead a security team, work as a C-suite executive, and shape an organization to be more defensible. So I wrote it. Scaling Security: The 7 Pillars of Elite Cybersecurity Teams philvenables.com/post/scalin…
6
29
4
150
33,345
Understanding system design helps you build better apps, even when vibe coding. Better architecture decisions. Better AI prompts. Fewer surprises when your app grows. So I built an interactive app to learn system design visually. designed-to-scale.vercel.app
171
590
46
5,516
294,576
Mental models that runs almost 93% of engineering
40
1,343
26
8,286
293,572
security nerds, this one is for you! The entire Security Engineering book by Ross Anderson is available online for free. An absolute goldmine of security knowledge. cl.cam.ac.uk/archive/rja14/b…
4
76
534
26,357
Many ppl in big tech are stuck in their level on some project for years. No growth, passed by over and over while others get promoted, etc. Nobody ever tells them why, just vague career molasses. Here's some advice if that's you: 1/ there are legible failures and ambient failures. Legible failures are specific-- e.g. you deploy smtg that causes an outage. Ambient failures are vague, everything is just kinda mediocre, nothing in particular to point to. Your work just isn't great. The reason nobody tells you anything concrete is that your failures are ambient. Very difficult to give feedback for those 2/ you need to decide what you want: job stability, money, or growth. Pick one. Ppl who are stuck tend to be unable to pick (which means they keep picking what they have now). Either accept it or make a different choice 3/ if you want job security, keep doing what you're doing, that's as secure as it's ever gonna get. If they didn't fire you by now that means you're doing something useful and firing you would be a pain in the neck. You're not safe in the long run because the world moves on and eventually the gap gets too large. But for the short term you're lindy, unlikely to get fired 4/ if you want more money you've got to interview, then either quit or come back with a better offer. You can't "negotiate" more money. Ppl will keep stringing you along because they know you're conservative and have no options, and there are others with far more risk appetite that need to be taken care of first. If you're unwilling to do that, fine, but then accept no raises 5/ if you want growth you need to leave your project and do smtg new. On the current thing you've built up too many assumptions, you have too many expectations, your situation empirically has not moved for a long time. This will be painful. It will feel like a setback. In the short term, it is. But in the long run you'll be a beginner again, get a kick in the butt, it will feel bad, but if you have your wits about you you'll soon surpass your former self. If you're not willing to do that, accept that you'll see no growth 6/ btw if you're a manager and want growth, become an IC for a while. All the best managers kinda dream of an IC role. Your barrier to growth is probably you not wanting to get your hands dirty. It will be painful and will feel like a setback. If you're unwilling to do that, fine, but then accept your situation 7/ all of this is very painful. If you choose staying/security you'll have to deal with the pain of resentment. If you choose interviewing/money you'll have to deal with the pain of prep and interview terror. If you choose to be a beginner again/growth, you'll have to deal with the pain of losing status (it's only in your head, but it's painful nonetheless). So it's pain in any direction, you only get to choose the kind of pain you face 8/ you've chosen doing nothing (and thus security/resentment) for years, which is good evidence you'll continue choosing that. The first step to any kind of change is seeing that you're doing it to yourself. You're choosing security/resentment because it's the least painful of the three options. But you can make a change tomorrow if you want 9/ there is no fourth easy choice, I'm sorry. I wish I could offer you one, but it doesn't exist /fin
12
16
4
591
36,720
For people who keep asking what to build in AI Security Engineering. ➣ Build your own Multi-Turn Jailbreak Fuzzer (stateful context grooming + Crescendo payloads + bypass scoring) ➣ Build your own MCP Tool Poisoning Shield (JSON-RPC sanitization + Pydantic validation + confused-deputy blocking) ➣ Build your own PII Unmasking Proxy (deterministic UUID swap + pre-flight masking + post-flight restoration) ➣ Build your own RAG Poison Detector (chunk hash verification + citation spoofing detection + shadow-context flagging) ➣ Build your own Grammar-Constrained Output Cage (CFG decoding + regex-enforced schemas + executable-code blocking) ➣ Build your own Delegated Agent OAuth Broker (scoped JWTs + time-limited tokens + least-privilege handoffs) ➣ Build your own Automated Red Team CI Gate (PyRIT payloads + jailbreak-rate thresholds + merge blocking) ➣ Build your own Semantic API Rate-Limiter (embedding similarity scoring + scrape-pattern detection + extraction blocking) ➣ Build your own Computer-Use Pixel Auditor (invisible UI detection + Set-of-Mark spoofing + click-target validation) ➣ Build your own Inter-Agent Sanitization Bus (cross-agent injection filtering + markdown stripping + trust-boundary enforcement) ➣ Build your own Vector Memory Integrity Scanner (sleeper-payload detection + episodic anomaly flagging + backdoor scrubbing) ➣ Build your own Cryptographic Trajectory Ledger (append-only hash chains + prompt/context/tool recording + tamper-proof replay) ➣ Build your own AI Supply Chain Scanner (hallucinated dependency detection + MCP server auditing + AI-generated SBOMs) ➣ Build your own Activation Patching Safety Dashboard (logit lens probing + deception-circuit discovery + sycophancy neutralization) ➣ Build your own Exploit Chain Teardown Engine (vulnerability reproduction + mitigation mapping + CVE-style disclosure) Pick 3. Build them from scratch. Document every decision. Most people trust the system prompt. Builders enforce constraints at the token level. Bookmark this. You'll need it.
12
21
267
11,973
Introduction to harness engineering for vulnerability research grahamhelton.github.io/harne…
6
106
1
919
44,412
Send this to whoever owns security at your company.
It's been a rough few weeks for security. I pulled together a guide of the highest-impact actions you can take today:
19
71
4
882
138,160
Michael Blackson penned a message and reflected on his own bullying experience while speaking on the death of fellow Ghanaian Javits Frimpong. The comedian also expressed plans to make appearances at high schools to talk to foreign kids and help them deal with bullying. “May God help us all to help deal with this. Amen.” Our thoughts are with Javits’ family 🙏🏽 ✍🏾#TSRStaffAS 📷:(@gettyimages)
171
1,163
138
10,871
980,447
The murder of a child is a pain no parent should have to endure, let alone from thousands of miles away. Javits Frimpong’s mother deserves to say goodbye to her son in person. I just got off the phone with President Trump and he has agreed to grant her visa immediately.  All of New York City is mourning this tragedy.
1,336
19,663
1,442
188,968
11,091,934
We got Microsoft Authenticator to hand over account tokens to a fake Teams app. The app needed zero Android permissions. It just claimed to be Teams, and Authenticator never checked who was actually asking. The result: full delegated access to mailbox, files, and calendar through Microsoft Graph. Full write-up by @ofekdavidlevin with code and PoC: enclave.ai/blog/authmask-we-…
Readers added context they thought people might want to know
Microsoft patched this (CVE-2026-80097) in Authenticator v6.2606.4246+. It requires a malicious app already on the device. No exploitation reported at advisory time. Update via Google Play. enclave.ai/blog/authmask-… msrc.microsoft.com/update-guide/v…
55
263
103
2,030
859,347
nerds, thank me later... systemdesign.io
121
1,299
81
15,579
2,552,483
I texted my children that we got the puppy they’ve been begging me for and sent them a link with pictures. 🐶When they clicked on the link, it took them to a webpage that said YOU’RE GROUNDED. They’ll be spending the weekend retaking our family’s cybersecurity training course.
177
6,534
209
111,162
1,380,345
messaging my wife a link containing proof of her husband cheating on her from an anonymous instagram account and then when she clicks on it texting her to let her know she failed the family phishing test and she has to retake my cybersecurity course
901
15,098
1,234
245,202
8,247,902
No. I think AI will eliminate entire classes of bugs, probably faster than most people expect. But the idea that we will stop shipping security vulnerabilities entirely by mid 2027 seems extremely unlikely. Take seL4, one of the best examples of formally verified software. You can mathematically prove code behaves according to its specification, but that doesn’t mean the entire system is secure. There are still assumptions about hardware, architecture, configuration, and everything else. Even when AI is designing the entire system, those assumptions still exist. And the same advances in AI that make systems more secure will also make attackers more capable. Cybersecurity isn’t like translation, where you can eventually approach some ceiling of performance. It’s an adversarial problem, and I don’t see that arms race ending any time soon.
Will the models stop writing code with bugs by mid 2027? teapawt.net/BVeiseh/status/2113389…
14
6
1
85
4,776
A guy built an AI that thinks it’s 1899, then tried to get it to beat Einstein to his most famous theories.
275
997
39
16,088
435,579