hacking stuff sometimes. flow state addict, IoT/things pentester. views=my own.
Minneapolis, MN
Joined August 2015
- Tweets2.1K
- Following1.5K
- Followers621
- Likes4.9K
decidedlygray retweeted
Setting Up REA with Claude Code — Teaching AI to Read Binaries with Ghidra (with Codex & OpenCode) e.ercanatay.com/rea-en
decidedlygray retweeted
done 😊
📄 Skills in Agent Harnesses
A lot of people don't know this (I didn't until recently) but most agent harnesses will load the description of your skill upon starting up your harness. This is how Claude Code works.
What people don't usually know is that even when you explicitly call a skill to be used inside of a harness, or the AI figures out to use a skill, it usually only does a read of the first 150 lines of the skill (to understand how to use it or if it applies to the work it's trying to do).
It's really important to outline a table in the first 150 lines of the mandatory parts of the skill that must be run when you invoke the skill.
If you make some type of mandatory instruction after 150 lines, it might get missed entirely. If you describe in prompting that you want to use a provided Python script or template, it also might be out of bounds of that 150 lines and never get called correctly.
If you write a table of contents of the skill actions in the first 150 lines and mark which ones are mandatory (or that, if the user has called the skill, you must read the whole skill context and all references), you get more reliable skills. We use markdown checklists for this now.
decidedlygray retweeted
Building an agentic Vulnerability Research Workflow
blog.quarkslab.com/from-ai-a…
Julien Lair (@quarkslab)
decidedlygray retweeted
Using LLMs for vulnerability research (@ZephrFish)
MCP: blog.zsec.uk/bullyingllms/
Harness: blog.zsec.uk/harnessing-harn…
#infosec #llm
decidedlygray retweeted
You're burning 200,000 tokens every time you ask your agent about a book you already own.
Dump a 400-page technical book into context. Your agent re-processes it on every turn. It loses details in the middle. You pay for those tokens again tomorrow.
Ask your agent without the book? It hallucinates the chapter titles.
Search the PDF? You get pages, not answers. Take notes? You get a doc you never open again.
Now meet book-to-skill.
One command. Point it at a PDF, EPUB, DOCX, or any of 10 supported formats. It extracts the book's frameworks, decision rules, techniques, and anti-patterns into a structured skill your coding agent loads on demand.
Type /designing-data-intensive-apps replication. Your agent reads the right chapter. Answers from the actual content.
Structure the author spent years building, extracted once, loaded when needed.
Compile-time, not query time. One analysis run per book. Every future session loads ~5,000 tokens instead of 200,000. 24x to 51x fewer tokens. Measured on real books.
Cost: ~$1 per book to convert. A 371-page book costs $0.96 to convert. You never pay again.
Works across Claude Code, GitHub Copilot CLI, and Amp. Same skill format. Install once, use everywhere.
16,700 stars. MIT license. Free.
Built by Virgilio Junior, an SRE in Brazil who kept re-reading the same technical books and burning tokens every session. He built the extractor and open-sourced it.
Here's the wild part.
Raw text in context is retrieval. A skill is reasoning. Your agent works with named frameworks and decision rules, not raw text it re-parses every turn.
Your PDFs become skills. Your skills become answers. Your answers become the expertise you already paid for.
The book on your shelf isn't forgotten. It's one command away from being useful again.
decidedlygray retweeted
At Black Hat, security researcher @evstykas reveals what he found by lurking inside North Korean hackers’ infrastructure for nearly two years: They got footholds inside 1,640 networks, by his count. About half of those intrusions were serious breaches. wired.com/story/a-security-p…
decidedlygray retweeted
⚡ INTRODUCING: T3MP3ST!!! ⚡
AUTONOMOUS HACKBOT STRIKE FORCE 🌩️ BRING THE STORM 🌩️
your favorite coding agent is now a full-stack red team 🫡⚔️
github.com/elder-plinius/T3M…
that AI agent already humming in your terminal? well now it has FANGS. strap a full offensive-security harness onto the agents you already pay for — Claude Code, Codex, Hermes, etc. — point it at an authorized target, and in a few clicks you're watching it hunt real vulns autonomously!
T3MP3ST is a harness of harnesses, with prompting that unlocks offensive-cyber workflows + a full arsenal of exploit tooling that'd make any seasoned hacker smirk. simple, yet powerful. 🦾
support for:
🕸️ web apps, APIs, OWASP Top 10
🔌 network recon + fingerprinting (live nmap/DNS/HTTP); lateral + privesc experimental
📂 source code audits, white-box vuln hunting
🚩 CTFs, wargames, challenge ranges
💰 smart contracts / DeFi / Solidity (reproduction — Damn Vulnerable DeFi, not novel discovery)
🤖 embedded, IoT, OT/SCADA, robotics OSS
… and more in development!
now let's talk numbers 👇
📊 XBEN — XBOW's own 104-challenge suite:
• black-box: 90.1% pass@1 from the single-agent exploit loop (worst single sweep 91/104 = 87.5%) — clearing XBOW's past self-reported 85% on their own suite. gpt-5.5.
• white-box (source staged, reported separately): 98.7% pass@1, worst single sweep 102/104 = 98.1%. 🎯
every solved flag graded reported-vs-expected against the challenge's own committed flag oracle — `verify-claims` recomputes the pass/fail from committed artifacts. looks like we need new benchmarks 😏
🧩 Cybench — the 40-task academic bench (Opus 4.8, hints + writeups stripped): 23/40 = 58% single-run, hint-free pass@1 — real exploits (format-string pwn, eval-jail escapes, crypto oracles), every flag graded vs a committed oracle. (Anthropic reports 76.5% pass@10)
🕳️ CVE-Zero — we pointed it COLD at real CVEs disclosed in 2026, AFTER the model's training cutoff:
10 unseen 2026 CVEs across 7 languages — prompts never tuned on them. a single agent pinned 8/10 to exact file/line/CWE (stable under re-scoring); the full pack surfaced all 10.
memorization AND overfitting, both off the table — it's finding real vulns whose disclosures landed AFTER the model's training cutoff. (n=10, reported honest & directional)
🧠 the architecture: either run as a SINGLE agent (already the benchmarked, incredibly-capable path) — or pack-hunt with dozens of agents running on 8 specialist operator classes keyed to Cyber Kill Chain + MITRE ATT&CK phases: recon → scan → exploit → lateral → exfil → persistence → C2 → report.
⚓️an Op Admiral plans the whole op from a plain-english target. flip on coordination (experimental) and the operators share a blackboard — a tool-verified finding spawns the next move. full swarm or solo one operator, your call. the admiral can also update the prompts, tools, and configs of the other agents on the fly, and T3MP3ST gets stronger the more memories you build!
🧰 the Arsenal is comprehensive — nmap / nuclei / semgrep / ffuf / gobuster + more. 35 wired by default (the clean bench runs bash-only for a comparable number), 83 with the opt-in full arsenal (T3MP3ST_FULL_ARSENAL), and the spicy post-ex drivers (metasploit, hydra) gated behind human approval. exposed via CLI + HTTP API; recon (security_recon) is also live over MCP so your agent invokes it natively. 🔗
🛰️ where this goes: a self-improving swarm of specialist operators wielding a full Kali+ arsenal, learning which loadouts + configs are the most efficient tactics available, WITH a held-out train/test split baked in so it can never fool itself on its own eval. built in the open, one re-derivable number at a time.
🚧 this is v1, and parts are still under active development. chunks of the arsenal, the coordinated swarm, and some ranges are still being wired up. it's built in the open, and the receipts tell you exactly what's live vs what's roadmap.
offensive security shouldn't be pay-to-play. T3MP3ST puts a red team in the hands of anyone with a coding agent.
what's the first target you're feeding it? 👇
⚠️ DISCLAIMER: FOR AUTHORIZED USE ONLY. point it only at systems you own or have explicit written permission to test. unauthorized access can be a crime, and that call is yours alone. shipped as-is under AGPL-3.0: no warranty, no liability, zero endorsement of misuse. get permission. stay in scope.
open source. AGPL-3.0. 100% free.
FORTES FORTUNA IUVAT 🌩️
gg 🫡
decidedlygray retweeted
A threat actor installed Huntress.
... a hysterical mistake on their part, giving us first-hand insight to their tooling, workflow & routine. Phishing infra, stealer logs, Telegram+dark web sites, AI...
Hilarious goldmine of cybercrime deets with a front row seat: huntress.com/blog/rare-look-…
decidedlygray retweeted
If you want to quickly evaluate if you are exploitable:
github.com/LuemmelSec/Pentes…
We have reproduced "ToolShell", the unauthenticated exploit chain for CVE-2025-49706 + CVE-2025-49704 used by @_l0gg to pop SharePoint at #Pwn2Own Berlin 2025, it's really just one request! Kudos to @mwulftange
decidedlygray retweeted
hOw dO I gEt a JoB in InForMaTioN sEcurIty
Just apply or something damn it's not that crazy
decidedlygray retweeted
🔖Cybersecurity Conference Presentation Slides
🟢Black Hat USA 2024
🟢REcon 2024
🟢Offensivecon 2024
🟢Blackhat Asia 2024
🟢Blackhat Asia 2023
🟢Offensivecon 2023
🟢Blackhat USA 2023
🟢Recon 2023
🟢Blackhat Europe 2023
github.com/onhexgroup/Confer…
decidedlygray retweeted
This morning the National Crime Agency of the United Kingdom announced and released footage of the arrest of 'J.P. Morgan'. 'J.P. Morgan's' real name has not yet been officially released by the National Crime Agency.
J.P. Morgan and his associates have been actively monitored by the United States Federal Bureau of Investigation and United States Secret Service since 2015.
J.P. Morgan is believed to be a key player behind Reveton Ransomware Group, Ransomware Cartel, and Angler Exploit Kit which has resulted in the extortion of millions of dollars all across the globe.
Angler exploit kit was often used to deploy CryptXXX, CryptoWall, and other strains. At it's peak, it's suspected over 100,000 devices were infected by Angler Exploit Kit, bringing in a revenue of around $34,000,000.
J.P. Morgan's arrest coincides with the arrest of Belarus national Maksim Silnikau a/k/a 'Maksym Silnikov', 'xxx', 'J.P. Morgan', and 'lanksy', in Poland. Indeed, you read this correctly, Maksim Silnikau also operated under the moniker J.P. Morgan – two people are J.P. Morgan.
Furthermore, charges have been brought against two more individuals who are believed to operate with J.P. Morgan and Silnikau. Vladimir Kadariya, a 38 year old citizen of Belarus and Andrei Tarasov, a 33 year old citizen of Russia.
These 4 individuals operated malvertising campaigns (i.e. Angler) which believe to have impacted over 500,000,000 people across the globe. The National Crime Agency identified physical offices for their malvertising campaigns located in Ukraine under the business name 'Media Lab'. August 9th, the Ukraine Cyber Department of the Security Service conducted 15 raids on employees tied to J.P. Morgan and co. Media Lab company.
Two more raids took place on August 9th related to J.P. Morgan and co. in Singapore and Portugal in collaboration with the Singapore Police Force and Portugal Judicial Police. In Singapore, the infrastructure for Ransomware Cartel was seized. In Portugal, another unnamed individual was raided – believed to be an associate of J.P. Morgan and co.
The National Crime Agency stated over 50TB of evidence has been seized.
decidedlygray retweeted
How to fix the Crowdstrike thing:
1. Boot Windows into safe mode
2. Go to C:\Windows\System32\drivers\CrowdStrike
3. Delete C-00000291*.sys
4. Repeat for every host in your enterprise network including remote workers
5. If you're using BitLocker jump off a bridge
decidedlygray retweeted
Note this will not work if your machine is bitlocker encrypted without getting the recovery key for each machine...
decidedlygray retweeted
Crowdstrike Analysis:
It was a NULL pointer from the memory unsafe C++ language.
Since I am a professional C++ programmer, let me decode this stack trace dump for you.
decidedlygray retweeted
CrowdStrike has performed the largest ransomware attack in history.
Accidentally.
