Sirack's Universe
Open in Telegram
Unfiltered thoughts. Brain dump and hot takes. Mostly about cyber security, malware analysis, reverse engineering, software dev, creative art and tennis (Rafa is the GOAT!) P.S. Opinions are my own; https://ismyapppwned.com https://sandboxbrief.com
Show moreThe country is not specifiedThe category is not specified
218
Subscribers
No data24 hours
+17 days
+330 days
Posts Archive
Royal decree by Google is killing Android's open ecosystem π€
I should give GrapheneOS a second chance
https://f-droid.org/2025/09/29/google-developer-registration-decree.html
Boom, finally published! It chronicles (and highly speculative at times :)) how LLMs can help with mobile security. I have disscussed few use cases
1. For code obfuscation
2. For reverse engineering
3. For device forensics
4. For SOC automation & reporting
https://www.sandboxbrief.com/opportunites-mtd-with-ai-agents/
I built my first mcp serve. It helps you iOS device forensics using iOS's sysdiagnose file. You can ask all kinds of interesting questions like
- how many apps were installed ? Past and present
- how many times was this device rebooted?
- find me suspicious apps
- what wifi was this device connected to, etc
"MIDROC is yours ; not mine" - Jesus
He is a shell of his former self π
Translation - I have lost control of my assets. You are on your own
Product idea - Use LLMs to obfuscate sensitive code in mobile apps. It's possible to get over hallucination if code has unit tests or AI can write one itself
Obviously itβs not straight forward like this
but definitely doable π
+1
One good use of LLMs is code obfuscation. In mobile apps, itβs very critical that code is protected from reverse engineering. There are many obfuscation techniques out there. Control flow obfuscation works by injecting dummy code that throws off possible attacker by making it difficult to read and understand the code.
See the screenshot above. On the left side is unobfuscated root detection code taken from https://github.com/scottyab/rootbeer/blob/master/rootbeerlib/src/main/java/com/scottyab/rootbeer/RootBeer.java#L182. On the right is code obfuscated using Claude Sonnet 4. Try to see if you can understand whatβt going on without executing it ;)
Here is the prompt used to generate the code https://claude.ai/share/1562ad16-695b-4748-90d3-47c2c91d2046
Sam Altman says that bots are making social media feel 'fake' | TechCrunch
https://techcrunch.com/2025/09/08/sam-altman-says-that-bots-are-making-social-media-feel-fake/
Repost from N/a
Framing Parliament.
Srilanka: we storm our Palace
Nepal: hold my beer
Yup.
I am planning to do more of teaching next year.
Some formal & mostly informal in the form of short form technical sessions and labs
Will provide deails soon
:)
