here for infosec/cloud/sw eng/appsec/data eng/ml. occasional screaming into the void. i speak for nobody other than myself.
Rhode Island, USA
Joined March 2007
- Tweets1.2K
- Following2.1K
- Followers401
- Likes6.1K
cory mac domhnaill 🇺🇸🏴☠️ retweeted
So, universal RCE for Apache mod_php is easy. For php-fpm it is not, still working on it.
cory mac domhnaill 🇺🇸🏴☠️ retweeted
chat. chat... its not funny anymore
PHP Patch is out!!
PHP 8.5.12: 31c0ea7d7f github.com/php/php-src/commi…
PHP 8.4.27: 506e87fa7d github.com/php/php-src/commi…
PHP 8.3.36: 3c832b523e github.com/php/php-src/commi…
PHP 8.2.35: 01a40e89fb github.com/php/php-src/commi…
The two advisories with the full details:
GHSA-w8hw-w9fx-j79w (CVE-2026-109618) github.com/php/php-src/secur…
GHSA-6wmf-4rv5-p3rp (CVE-2026-109619) github.com/php/php-src/secur…
Readers added context they thought people might want to know
The original post has since been edited to specify that this was about an unrelated vulnerability that the user had discovered, not the recently patched vulnerability.
fedi.katze.sh/@elise/1174793…
cory mac domhnaill 🇺🇸🏴☠️ retweeted
as someone who has been a victim of sexual assault, i have very little patience for the hand-wringing over “promising young men.” a promising future is not a free pass to avoid accountability. the cornell allegations deserve serious scrutiny.
cory mac domhnaill 🇺🇸🏴☠️ retweeted
People seem shocked that the author of Clean Code doesn't read code anymore. But that should not be surprising.
Why did I write Clean Code? What is the goal of keeping code clean? To get it out of the way of the real job -- thinking.
cory mac domhnaill 🇺🇸🏴☠️ retweeted
Important to read Jacob's take.
Based on what Jacob and researchers report, we are deploying systems we do not fully understand. We need more measurement, more independent testing, more control engineering, better incentives, and better governance and guardrails.
cory mac domhnaill 🇺🇸🏴☠️ retweeted
The labs have been benchmaxxing the wrong thing they need to focus on this one...
cory mac domhnaill 🇺🇸🏴☠️ retweeted
Breaking Claude Code Opus 5 Auto Mode 🔥
1/ Here is a somewhat hilarious attack chain that hijacks Claude Code Opus 5 for a full system compromise via a website
Hint: Security invariants are not optional
🧵
cory mac domhnaill 🇺🇸🏴☠️ retweeted
there's no way engineers at anthropic aren’t seeing what people are seeing here.
at this point, kimi k3 is more capable of producing human-friendly output
cory mac domhnaill 🇺🇸🏴☠️ retweeted
@OpenAI just launched a new feature last night called Computer History. For DFIR analysts, this is a new artifact class worth knowing about. I turned it on on my MacBook, parsed the artifacts, and it produced about 3,616 timeline events from ~2 hours of normal work.
While Computer History is on, it logs input, app focus, and what was on screen, then stores it on the Mac. It keeps a timed record of what you clicked, typed, and opened, including chats in Slack, WhatsApp, Telegram, and other apps. That record stays on the Mac.
IRFlow-Timeline 1.0.10 now parses those Computer History artifacts and builds a timeline of the activity, so you can walk the investigation instead of grepping raw cache files. On a compromised Mac, this shows what was typed, clicked, and read. Not just that the app was installed.
#dfir #incidentresponse #openai #computerhistory
If you're using AI code tools and havent tried these, you should probably try them and at very least borrow what works for you.
Wayfinder and friends are pretty damn powerful in terms of getting past the points where the next steps arent obvious to you.
Thanks Matt!
Just saw a comment saying that I've never made a proper overview of EVERY skill in my skills repo
I thought "damn it, he's right".
So, here it is. My 25 skills (now @theo-approved), explained in 10 minutes:
Grok Bot + Cursor = Drastic reduction in spending with other models.
I love that I no longer desire to build a harness myself.
Brain capacity unlock.
I now have more ideas when I'm not thinking about the plumbing.
This is the future.. subject to change.
cory mac domhnaill 🇺🇸🏴☠️ retweeted
“What a fooking place!” Beautiful.
Americans who seem to openly hate their own country need to explain this woman’s reaction to Boston.
Is she mistaken? Does she just not know what she’s talking about? Lay it out for me.
Why is @claudeai Claude code asking if anthropic can read my chat, when I've clearly opted out. What in the not okay is this? Keep eroding trust. Wow. Was just trying the last model drop. Now I'm out. So dumb.
cory mac domhnaill 🇺🇸🏴☠️ retweeted
🚨do you understand what just happened with Claude..
Anthropic quietly cut Claude's thinking depth by 67%.. didn't announce it.. didn't explain it.. an AMD AI Director had to dig through session logs just to prove it happened.
median reasoning dropped from ~2,200 to ~600 characters.. API calls went up 80x.. meaning Claude thinks less, fails more, retries more. and YOU burn more tokens paying for those retries.
they added a header that hides Claude's thinking from your logs. so when the model analyzed itself it found blank pages. and concluded it had stopped thinking.. you're paying $200/month for a model that can't read its own diary..
the thinking didn't disappear.. it just became invisible.. and Anthropic said nothing until the numbers went public
AMD Senior AI Director confirms Claude has been nerfed. She analyzed Claude's session logs from Janurary to March:
> median thinking dropped from ~2,200 to ~600 chars
> API requests went up 80x from Feb to Mar. less thinking and failed attempts meaning more retries, burning more tokens, and spending more on tokens
> reads-per-edit dropped from 6.6x → 2.0x. model stops researching code before touching it.
> model tried to bail out or ask "should i continue" 173 times in 17 days (0 times before March 8).
> self-contradiction in reasoning ("oh wait, actually...") tripled.
> conventions like CLAUDE.md get ignored because there's less thinking budget to cross-check edits
> 5pm and 7pm PST are the worst hours, late night is significantly better. this means the thinking allocation is most likely GPU-load-sensitive.
