Do It by Code
64 subscribers
771 photos
114 videos
16 files
1.35K links
We uhhhhh... do things by coding them.

"This is wild, multi-agent coordination, clearly infrastructure hacking. We should not."
Download Telegram
# How IDA Pro's License System Was Defeated With a 1-Byte Patch

## How IDA's licensing works

IDA Pro validates licenses using RSA signature verification. License files (`.hexlic`) are JSON documents containing license details, signed with Hex-Rays' private key. The binary contains a hardcoded 1024-bit RSA public modulus to verify signatures:


signature ^ e mod N == SHA256(license_payload) ?


Without the private key, you can't forge a valid signature. Breaking the original 1024-bit RSA key is computationally infeasible. But the attacker didn't break anything — they replaced the lock.

## The attack: modulus substitution

The attacker modified one byte of the 128-byte RSA modulus embedded in the binary:


Original: ED FD 42 [5C] F9 78 54 6E ... (124 more identical bytes)
Patched: ED FD 42 [CB] F9 78 54 6E ...


The critical insight: the resulting number is prime. When N' is prime, phi(N') = N' - 1 is trivially known, so the private key is just:


d = e^(-1) mod (N' - 1)


A single modular inverse — effectively free to compute. No factoring, no brute force, no heavy computation. Standard RSA's security relies on N being a product of two large primes whose factorization is secret. A prime modulus has no trapdoor — there's nothing to "factor," so anyone can derive the private key.

## How hard was this to find?

Not hard at all. For a random 1024-bit odd number, the probability of being prime is roughly 1/355. Scanning all 255 alternative byte values at one position takes under 400ms. Scanning every one-byte mutation across all 128 byte positions (32,640 candidates) takes under 45 seconds and yields 86 usable primes.

The entire computational cost of this attack is less than a minute on a single machine.

## What the keygen does

1. Patches the modulus — a Python bytes.replace() swapping 6 bytes (only 1 actually changes), targeting libida.dylib, libida.so, ida.dll and their 32-bit variants
2. Generates a `.hexlic` license — JSON with all decompiler add-ons enabled (x86, ARM, MIPS, PPC, RISC-V, ARC, V850), Teams, Lumina, 999 seats, valid for 3-10 years
3. Signs it with the derived private key — SHA-256 of the payload, packed into a 128-byte buffer with a custom padding scheme, RSA-signed with the attacker's d

IDA verifies the signature against the patched modulus and sees a perfectly valid license.

## Why it can't really be fixed locally

The verification code itself is fine — the flaw is architectural. The binary trusts an embedded public key, and the attacker can edit the binary. Once the trust anchor is replaced, the cryptographic scheme is irrelevant. Larger keys don't help (twice the bytes but half the prime density — roughly cancels out). Better padding doesn't help. Self-integrity checks can be patched out the same way.

The only robust defense is moving the root of trust outside the patchable binary: online activation, server-side feature gates, or hardware-backed attestation. But IDA Pro is an offline-first desktop tool whose users are world-class at patching binaries — which is why this cat-and-mouse game has been going on for 20+ years.
LEDGER SUPPLY CHAIN ATTACK MAY HAVE BEEN EXPOSED BEFORE MASS WALLET DRAIN

Former Mt. Gox CEO Mark Karpelès may have inadvertently triggered a coordinated wallet drain after publicly exposing physically modified Ledger devices containing hidden surveillance implants.

On October 8, Karpelès warned that counterfeit or tampered Ledger devices were being sold with concealed SIM cards capable of transmitting stolen seed phrases.

The implants reportedly intercept data sent to the device’s screen, allowing attackers to capture recovery phrases displayed during wallet setup.

Crucially, these devices can reportedly pass Ledger’s official Genuine Check because the original secure element remains intact. The verification process cannot detect unauthorized physical modifications.

One theory now emerging is that attackers had been harvesting seed phrases from compromised devices, waiting to drain multiple wallets simultaneously.

Karpelès’ warning reportedly attracted roughly 90,000 views, potentially alerting the attackers that their operation had been exposed and prompting them to move stolen funds.

Karpelès says one compromised Ledger he examined originated in Malaysia, arrived in flawless shrink-wrap packaging, and contained an implant cleverly concealed beneath the screen.

The suspected attack vector is a hardware supply chain compromise, not a cryptographic vulnerability.

The connection between Karpelès’ warning and the subsequent wallet drains remains unconfirmed, as does whether the same implants were responsible.

If established, the incident would demonstrate how compromised hardware can undermine self-custody even when a device passes manufacturer authentication.

Posted by Bitcoin News, 4 hour ago


so here is a TLDR:
some people bought lots of ledger wallets (they are physical hardware wallets)
then put sim cards in them so they can transmit the seeds to themselves
then resold them
and waited for the right time to drain all
😭1
I'm gonna be straight up honest:

"The Ripple team awarded us the maximum $250,000 bounty offered by their bug bounty program. To our knowledge, that's the largest ever paid out for a vulnerability discovered entirely by an AI agent."

This is laughable for a bug like this, and the signal it sends is dangerous. The bounty should have been at least one more zero. AT LEAST.

Ripple sits on an infinite amount of cash thanks to the monthly escrow release, and the best reward (assuming there are no others we know about it) they could come up with is 250k for something that could have wiped out their entity's existence.

Biggest aura loss ever; this is outright pathetic.

Posted by Daniel "CEO of the XRPL" Keller, 12 hours ago
Do It by Code
I'm gonna be straight up honest: "The Ripple team awarded us the maximum $250,000 bounty offered by their bug bounty program. To our knowledge, that's the largest ever paid out for a vulnerability discovered entirely by an AI agent." This is laughable for…
tldr a team with their AI agent found a critical bug in XRP that let anyone mint a gigallion tokens instantly

"The bug we found allowed any unauthenticated attacker to create 18 trillion XRP with a single transaction, 184 times the total supply."

"Attackers can inflate the XRP supply by trillions by triggering an integer overflow in offer processing. This catastrophic flaw bypasses inflation checks and severely compromises network integrity."

PoC
Do not install Zod 4.6.6 – it is infected
Article, Comments

update: they went lucky; their publish was blocked before normal registry distribution. npm's pre-publication security control prevented it from becoming normally installable from the registry.

the screenshot is from d821685
malware developers are also becoming lazy, at least try to hide your payload fetching
they used to obfuscate it or something

in short: it calls an eth contract and loads/injects the payload
😱2
Linus Torvalds Working On A Hefty Linux Kernel Improvement With Claude Code

Linus Torvalds is using Claude Code to develop a significant Linux kernel improvement on a new branch named noref-filemap-read-with-llm. The project aims to optimize memory management by allowing page cache reads without folio references, which resolves contention issues during small multi-threaded file reads. Torvalds confirmed that the LLM assisted with coding, testing, and drafting the patch series, noting that the changes deliver substantial performance gains. The patches are currently pending for potential inclusion in the upcoming Linux 7.4 merge window.


- commit: 4c0f369
- branch: noref-filemap-read-with-llm (not merged yet)
Introducing MoonMapBench

Can AI predict the elevation of 16,200 moon coordinates?

Opus 5.5 on average is off by 1,192 m (3,911 ft).
GPT 6 Astra is worse with 1,490 m (4,888 ft).
🧵

Posted by Ben Rank, 45 minutes ago

https://github.com/aisa-group/MoonMapBench
Re: Hackers activate automatic defrosting in supermarket fridges across Europe

Was just talking to a lab tech at a hospital.

It seems that some of the newer lab machines require internet access because you can't run lab tests unless you have a paid subscription to use the machine. No credits on the subscription, no tests.

I mention this b/c a lot of commenters here are saying "but why would you ever hook up a freezer to the internet??"

I find that type of comment fascinating as many of the people on HN worked for firms that built the technology required to use internet based subscription models.

alexpotato, 9 hours ago
11 year old unsoundness bugs found in Rust borrow checker
(struct's lifetime requirements can be bypassed via covariant field assignment)
Article, Comments
Do It by Code
Re: Hackers activate automatic defrosting in supermarket fridges across Europe Was just talking to a lab tech at a hospital. It seems that some of the newer lab machines require internet access because you can't run lab tests unless you have a paid subscription…
Regarding the Danfoss refrigeration systems hacked in France:

The username is admin, and the password automatically changes every day to... the current date.

Yes, today's date.

The French supermarkets affected in today's attacks had exposed all its refrigeration controllers directly to the internet.

Every single device reportedly used the same password: today's date...

There's also a documented authentication bypass: sending a POST request to /xml.cgi can bypass the login entirely.

Via: GossiTheDog

Posted by International Cyber Digest, 17 hours ago
we got some critical vulns in chrome after some time:
(the fixes for these CVEs are shipped in the October 13, 2026 desktop update: Stable 155.0.8059.79/.80 and Extended Stable 152.0.7977.167)

(sorted by importance in my personal opinion)

use-after-free RCE (sandbox escape):
- CVE-2026-110136: PDF subsystem (147 for inside of the sandbox)
- CVE-2026-110148: SVG subsystem
- CVE-2026-110149: Browser subsystem
- CVE-2026-110151: Media subsystem (153 for inside the sandbox)
- CVE-2026-110161: Network subsystem
- CVE-2026-110165: Core subsystem
- CVE-2026-110142: Filesystem subsystem (this one is rated Medium, idk why. but yes it's sandbox escape)

buffer overflow RCE (sandbox escape):
- CVE-2026-110152: ANGLE subsystem (apparently Android-only?)


other cool ones:

CVE-2026-110143
Use of uninitialized variable in ThirdParty lmza_sdk: read memory inside the sandbox via a crafted file.

CVE-2026-110145
Use after free in Passwords subsystem allowed a remote attacker who had compromised the renderer process to execute arbitrary code outside the sandbox via a crafted HTML page. (Chromium security severity: High)


- Release blog
Sorry @alexandr_wang but I think your new model is a bit color blind

Posted by Jonathan Chang, 4 hours ago
Claude reveals ECDSA curve origins
Article, Comments
https://words.filippo.io/seeds-solved/

This isn’t an ECDSA break: it explains how the public curve parameters were chosen, making deliberate manipulation less plausible. Some details remain unexplained, including why certain valid candidate curves were skipped.
The finding: the mysterious seeds behind P-192, P-224, P-256, P-384 and P-521 came from an old C-library pseudorandom generator initialized with Unix timestamps from 1997–1998, rather than hashes of supposedly forgotten phrases.
This media is not supported in your browser
VIEW IN TELEGRAM
DeepSeek V4.1F-1017 has built this in like 5 minutes.
(Since then it's autistically rabbitholing into minor details)
These are unambiguously better than Opus's.
It's a good deal stronger than V4.1.

Posted by Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞), 3 hour ago

DeepSeek-v4.1-Flash-1017 on huggingface
Article, Comments
🚨 Confirmed. OpenCode hacked. Keys stolen.

Telling customers to revoke.

Posted by Matt Johansen, 2 hours ago
Congratulations to OpenAI on this great model release!

Posted by Lisan al Gaib, 30 minutes ago
CVE-2026-99999: PwnEscape: linux kernel full Guest-to-Host KVM Escape
(score: 9.3 critical)
blog post
vercel blog post

ipv4: If a packet with bad checksum metadata gets into the ipv4 stack, skb_checksum_help can corrupt the network header.

Affected versions:
All versions above 2.6.27 unless they got a patch applied to them, many branches already got the patch backported to them:
• 5.10: through 5.10.271 → fixed in 5.10.272
• 5.15: through 5.15.222 → fixed in 5.15.223
• 6.1: through 6.1.189 → fixed in 6.1.190
• 6.6: through 6.6.158 → fixed in 6.6.159
• 6.12: through 6.12.112 → fixed in 6.12.113
• 6.18: through 6.18.56 → fixed in 6.18.57
• 7.2: through 7.2.10 → fixed in 7.2.11
• 7.3 release candidates: through 7.3-rc6 → fixed in 7.3-rc7



Mitigation
==========

The Linux kernel CVE team recommends that you update to the latest
stable kernel version for this, and many other bugfixes. Individual
changes are never tested alone, but rather are part of a larger kernel
release. Cherry-picking individual commits is not recommended or
supported by the Linux kernel community at all.

If however, updating to the latest release is impossible, the individual changes to resolve this issue can be found at these commits:
- a464d95
- 0fab003
- 94524b7
- a65f69e
- 32a8b0a
- f3e835a
- c3e3346
- 6785011
The Intel 8087 chip: floating-point addition and subtraction

Intel 8087 chip introduced in 1980 significantly improved floating-point operations, with a speedup of almost 100 times over the 8086 processor. The microcode implementation of addition and subtraction involves complex logic to handle special cases such as infinities, zeros, and NaN values.