@aaronott.bsky.social
Spent some time digging through the big AI security reports from 2025.
Some expected themes, some wild surprises, including a zero-click prompt injection against Microsoft 365.

www.ado.im/posts/ai-sec...

If you’re working in AI or AppSec, curious what trends you’re seeing.
#ai-security
AI Security in 2025: What I Learned This Weekend, and What I'll Be Watching for in 2026 | Aaron Ott
A weekend deep dive into three major AI security reports from 2025, what actually broke in the real world, and where security teams should focus their attention in 2026.
www.ado.im
December 14, 2025 at 11:53 PM
Claude 4.5 ran a 15-min pentest on my Kali lab — 21 real vulns found, but it missed XSS and logic bugs. Great for early checks and demos; don’t rely on it alone.

Read: ado.im/posts/ai-pen...

#AIsecurity #pentesting
October 3, 2025 at 9:57 PM
Built a lightweight analytics stack by pairing ChatGPT with Claude Code.

ChatGPT as architect for first drafts.
Claude Code as repo surgeon for precise edits.

Full write-up with prompts and guardrails:
ado.im/posts/coding...

#AI #ChatGPT #ClaudeCode
Coding With Chatgpt vs Claude Code | Aaron Ott
I built a lightweight analytics stack: ChatGPT for fast scaffolding, Claude Code for surgical repo edits—steal the prompts, guardrails, and NGINX lessons.
ado.im
September 14, 2025 at 6:44 PM