PostgreSQL pgcrypto heap buffer overflow executes arbitrary code
CVE: CVE-2026-2005
PT ID: PT-2026-7845
Vendor: PostgreSQL
Product: PostgreSQL
CVSS: 8.8
Credits: Team Xint Code
Description:
Heap buffer overflow in PostgreSQL pgcrypto allows a ciphertext provider to execute arbitrary code as the operating system user running the database. Versions before PostgreSQL 18.2, 17.8, 16.12, 15.16, and 14.21 are affected.
References:
• dbugs.ptsecurity.com/vulnera…
• postgresql.org/support/secur…
PoC/Exploit: github.com/var77/CVE-2026-20…
#dbugs_vuln
2
11
521
Old Fr pervert retweeted
11 Apr 2025
How I like to treat my black women colleagues. Hope you are getting treated right at your work ladies.
5
78
825
238,591
CVE-2026-31431 ("Copy Fail") is a deterministic local privilege escalation (LPE) vulnerability in the Linux kernel that has been sitting in production kernels since 2017. Publicly disclosed on April 29, 2026 by Theori/Xint researchers, and added to CISA's Known Exploited…
1
52
Jun 12
Fuck of terrorist cunt. Get my money back of Zelensky you xint he stile 200billion fuck off Cunt Cunt Cunt
Jun 11
🚨HORROR IN IRELAND
Two Migrant Invaders beat an Irishman HALF TO DEATH outside of a hotel in broad daylight.
Locals had to LEAVE THEIR CARS and RISK THEIR LIVES to stop the murder.
This is what ‘diversity’ looks like in Ireland.
When TF does this STOP?
11
Jun 11
🚨 Happy to announce that we have been named Best Autonomous Pentesting Platform by @TheHackersNews:
"We're pleased to recognize Xint as the winner of the Automated Pentesting award in the 2026 Cybersecurity Stars Awards. Xint uses AI to scan application code and live systems for vulnerabilities, aiming to find serious bugs quickly while keeping false positives low. The team's deep experience in offensive security research informs their work, and we appreciate their contributions to strengthening the security field."
This achievement was evaluated and selected by an independent panel of judges appointed by The Hacker News, recognizing contribution, innovation, and impact within the cybersecurity industry:
awards.thehackernews.com/win…
1
2
20
7,735
Doesn't Mythos instantly make most of AIxCC obsolete?
We suspect Xint might be the only team that still has some edge, probably not because of their harness or AI, more because of their domain expertise. PPP has been dominating the CTF world for years; that has to count for something.
MDASH, on the other hand, looks like pure hype: a sophisticated-sounding framework built by a team that does not have a strong model. We could be wrong, but it is difficult to see where the advantage would come from.
In general, we remain skeptical that a harness can provide a durable moat. Again, if you think your vibe-coded harness is what keeps you ahead of AI, just wait for the next model release.
The things that compound are expertise, intuition, and hard-won domain knowledge. Tooling is the part everyone else can copy by Monday.
Of course, if you must do something to justify your existence, then building a harness makes perfect sense.
3
12
781
May 31
A 9-year-old Linux kernel local privilege escalation bug — with a full exploit. CVE-2026-31431 hits Ubuntu 24.04, Amazon Linux 2023, RHEL 10.1, and SUSE 16. Theori's Xint Code team found it and published the PoC. Go patch your stuff. ⭐ 3.9K #Security #Linux
github.com/theori-io/copy-fa…
Follow for daily dev finds 🔔
2
530
May 29
Check out the full interview with Xint researchers @tjbecker and @MountainRo53 on Google @wiz_io podcast. They cover:
🪲 How Xint's custom LLM harness uncovered CopyFail, a privilege escalation bug affecting almost every Linux machine since 2017.
⏱️ The harsh reality of vulnerability disclosure in the AI era—why 90 days is too long when models can weaponize patches instantly.
🤖 The evolution of AI agents in security, from the DARPA AI Cyber Challenge to Claude 3.5 Sonnet and Mythos.
💣 Why false positives are actually just undocumented zero-days.
youtu.be/ZX1Ot50kYcU?si=C6R_…
1
2
12
2,810
May 26
The first question we get when talking with CISOs is if Xint/AI replaces human cyber experts because LLMs are so good at bug discovery.
Short answer: no
Long answer: still no...and in fact human expertise is more important with the intersection of AI generating more insecure code faster than humans can review and AI-enabled attackers finding and exploiting vulns faster than teams can patch - but these human offensive security experts need the right AI platform
xint.io/blog/ai-cybersecurit…
2
10
1,313
Maddison taking the reins from Coady for Expected Interviews (xInt) per Season 📈
May 26
"At times we didn't deserve the support that they gave us"
James Maddison reflects on a difficult season for Tottenham Hotspur after they avoided relegation from the Premier League 💬
9
2
293
15,865
May 20
Because of the way we architected Xint, we are able to optimize the models we use for each agent workflow - that way our customers can take advantage of the improvements of each new model without having to rebuild their pipeline
May 19
Quick @xint_official benchmark of @GoogleDeepMind Gemini 3.5 Flash (vs 3.0): TP recall is the same, FP precision is noticeably better. 3.1 Pro still the clear winner.
3
476
May 18
'When Anthropic announced Claude Mythos Preview and said it was too dangerous to release publicly, prompting shock and alarm across the business community, the former @DARPA competitors mostly just shrugged.
“It’s very cool,” [Xint researcher] @tylerni7 said, but “this is the world that we’ve been living in for a while now.”
Still, the new publicity surrounding AI vulnerability detection could benefit the teams behind the open-source systems. “It leads people to find out that, ‘Oh, this is a thing that my company should be worried about,’” Nighswander said.'
Great writeup from @CyberSecDive on DARPA's post-AIxCC efforts to take the innovation from this contest and apply it to securing critical open source and commercial digital infrastructure.
cybersecuritydive.com/news/a…
1
7
34
7,020
May 16
Shipped a big update to xint - my X intelligence CLI for terminals & AI agents.
🪙 Truthful Grok credit onboarding (X Premium ≠ API access - most guides get this wrong)
💸 24h cache on follower diffs: $50 → $0 on repeat runs
🎯 --budget cheap|balanced|max model routing for Grok 4.3
👁 --dry-run cost previews before any API call
📈 xint costs forecast projects end-of-month spend
⚡️ ~60-80% monthly cost reduction for heavy users
Both the TypeScript and Rust binaries got it. 231 tests passing.
github.com/0xNyk/xint
2
7
11
916
May 13
The speed of change in security research is stunning: when we won @wiz_io's ZDC in December, we were the only team using LLMs with zero human intervention. 3 scans with Xint Code was enough to win.
6 months later and the number of LLM entrants has completely overwhelmed Pwn2Own - and this isn't AI slop - these are legit 0days.
What does this mean for open source project maintainers or product security teams charged with securing massive codebases or thousands of public-facing apps, servers, applications, and APIs?
‼️🚨 Pwn2Own Berlin 2026 just hit a wall. For the first time in 19-years, ZDI rejected dozens of working zero-day RCE submissions because organizers ran out of contest slots.
Rejected hackers are now going public with PoC demos and direct vendor disclosures, breaking Pwn2Own's usual secrecy.
▪️ AI surfaces a massive wave of 0-day RCEs.
▪️ Submissions overwhelm ZDI past max capacity.
▪️ Slots run out. Researchers with working chains get rejected.
▪️ "Revenge disclosures" begin. ← we are here.
Confirmed casualties so far:
▪️ @xchglabs : 86 vulnerabilities prepared (PyTorch, NVIDIA, Linux KVM, Oracle, Docker, Ollama, Chroma, LiteLLM, llama.cpp). All rejected. Now reporting directly to vendors with writeups dropping as patches land.
▪️ @ggwhyp : full-chain Firefox RCE on Windows. Rejected. Publicly demoed (HTML page → cmd.exe → calc.exe). Responsibly disclosed to Mozilla.
▪️ @yunsu_dev : working RCE chain, rejected. Submitting elsewhere.
▪️ @ryotkak : tried to register for 3 weeks. ZDI confirmed "at maximum capacity, can't add extra contest days." Considered canceling flight and hotel.
▪️ @anzuukino2802 : Claude Code RCE PoC. Rejected.
▪️ @desckimh : 0-day RCEs in Ollama and LM Studio. Rejected.
Reported impact: a community-estimated 150 researchers tried to register. Accepted contestants are now being warned about collisions. Rejected vulnerabilities going to bug bounty programs may trigger pre-event patches that invalidate the work of those who got in.
ZDI has not publicly addressed the capacity issue. The event still runs May 14-16 in Berlin.
1
6
28
4,421
May 12
Here is a brief overview of two kernel-level vulnerabilities uncovered by Xint Code in MacOS, iOS and iPadOS and now patched: CVE-2026-28972 and CVE-2026-28986
xint.io/blog/kernel-vulnerab…
6
48
3,462
May 11
What assumptions are you making about how your data is being used and stored when you use LLMs to find bugs in your code? Zero data retention is not the default even for Enterprise agreements. See how Xint provides double protection both in terms of your data store agreements with us as well as our data store agreements with the frontier models.
xint.io/blog/zero-data-reten…
1
6
518
May 11
Copy Fail: 732 Bytes to Root on Every Major Linux Distribution.
Xint Code disclosed CVE-2026-31431, an authencesn scratch-write bug chaining AF_ALG splice() into a 4-byte page cache write. A 732-byte PoC gets root on Ubuntu, Amazon Linux, RHEL, SUSE.
xint.io/blog/copy-fail-linux…
14
14,254