@lostsec_ Thanx for the reminder of an old vuln in 2026. 😉
Your payload did not work in my tested target, but the poc of snyk-labs did its job on very old version 1.x 🔥
Will search for more projects with pdf.js 💥
@intigriti seen a lot as default in internal test stages, sometimes after deployment to prod, too.
But in Germany we are safe, because we have the "Hackerparagraph" § 202a StGB, were it's forbidden to use this login. You don't go over start, do not get 4000€, but directly into jail.
@intigriti I'm going to the corner of shame now. I already read about jsonp callbacks and simply did not test it.
I interacted with the bot's browser a lot, but funny script part to internal was blocked by SOP.
I was on the red herring with outdated chrome 129 version.
But! I learned a lot.
@chux13786509 Almost 6 months are over.
I’ve actually been hooked from time to time and have read and learned a lot. For example, the PG simple query protocol, but apparently nothing that applies to this code snippet.
Please explain, I really don't want to die a fool.
@intigriti 5 seconds for .htaccess and .htpasswd. It was where I meet John first time 🤣
But it was systemic behavior in past, ~ 2000
Back in the day, I didn't think in terms of vulnerabilities, but rather in terms of bugs and misused "features".
@_aircorridor@three_cube ❣️This exploit works fine on unpatched Ubuntu based systems.
My production RHEL systems don't have installed rpmbuild, so attack surface is not given.
Shift right with right shit by shit left.
🤭
Happens if a project tries to move security responsibility to the end, from left to right and is no DevSec anymore.
But only one character in the word "shift" was lost.🤣
New week, new supply chain fun.
A package.json with entries for version pinning, e.g.
"axios": "^1.5.1"
vs.
"axios": "~1.5.1"
vs.
"axios": "1.5.1"
and `npm install` vs `npm ci` vs. `yarn add` but without yarn.lock
Can you see the problem?
A lot of projects had "fun" today.
@AquaSecTeam To be precise, an internal CI/CD merge request from renovate bot, simply to update trivy version.
It was not only bad, it was a nice workload, learned some more on top.
@aquasecteam is doing a great job with trivy.
Usually, you can use the results to demonstrate the vulnerability fairly quickly using existing exploits.
But today, a simple renovate merge request compromised several projects. Supply chain attacks everywhere
https://t.co/CkeAOilSWG
@chux13786509 get public key from /.well-known endpoint, create a jwt with algo HS256 and sign with publ key.
the verify jwt will try to decode not only with RS256, but with HS256 as well. bingo!
Can I get now the solution for your golang challenge?🙏
My brain burns
@chux13786509 Is glitch on this client code level only or maybe on PG server with some bad pg_rewrite custom rewriter?
I hate you for the challenge because I don't see the problem between the lines.
I love you for forcing me to dive deeper into Golang, side-channel attacks via Sort and PG.
@chen9918b@msftsecresponse is it that simple I reproduced and with user interaction?
I need to click on the link in notepad md view and while unpatched version opens the calc.exe directly, the patched version asks me because the link could be unsafe.