@malwareunicorn

Mostly on @malwareunicorn.bsky.social @Straikerai. Ex-Microsoft. Ex-Meta.

 San Francisco, CA
malwareunicorn.org
Joined February 2014
  • Tweets2.1K
  • Following6
  • Followers166K
322Media
Malware Unicorn retweeted
Straiker
@straikeraiOct 21
New research: #CyberCartel's #CARLIY0S runs a multi-model #AI attack chain, splitting recon, exploit development, and data exfiltration across #Claude, #Grok, and #Kimi. Operators used "authorized red team" prompt framing to slip past model guardrails. The first AiPT we've documented from inside the operation.
3
7
4
12
1,586
The rare cfp badge
11
16
443
21,150
I stopped by the malware village for this alleged Ida pro swag.
1
1
21
7,600
Playing with my new giant portal badge ❤️
3
4
76
8,522
Malware Unicorn retweeted
Straiker
@straikeraiJul 19
"Defenders prepared for hostile users and rogue models. Here the model obeys, the user is innocent and the context is the weapon." straiker.ai/report/threat-re…

The State of AI Agent Attacks: Straiker Threat Report Vol 1

New data on 1,700+ successful attacks against AI agents. 36% reached code execution, 91% ended in silent data theft. Read the full threat report by Straiker.

straiker.ai
4
5
1,629
Malware Unicorn retweeted
blasty
@bl4styJun 17
1
50
4,884
Me: I went to the dentist. my 5yr old kid: Did you get a computer? me: ? kid: You’re big, and kids are small and get toys at the dentist. me: lol why? kid: bc you hack things. Can’t make this stuff up.
4
1
123
9,971
Malware Unicorn retweeted
Jinjing Liang
@JinjingLiangJun 13
Replying to @AnthropicAI
The state of things:
74
503
48
10,172
782,244
I only see cybersecurity job security.
fidexCode
@fidexcodeJun 8
Agreed?
5
5
72
9,215
Malware Unicorn retweeted
Straiker
@straikeraiJun 8
🎮 @straikerai just launched the world’s first fully playable video game CTF for agentic AI security. Chapter 1 includes 3 levels: Jailbreaking Command Injection Memory Poisoning It’s a full cyberpunk world where every level teaches you how AI agents can be tricked, manipulated, and broken. Follow us for each new chapter release. Play it: grid-city.straiker.ai |#videogame #infosec #ctf
1
7
10
2,458
Malware Unicorn retweeted
gabsmashh
@gabsmashhJun 7
47
893
23
13,211
272,524
Malware Unicorn retweeted
b1ack0wl@b1ack0wlMay 28
okay, now tell them that their vulnerability doesn't qualify for a bounty, but then patch it in the next release.
16
163
9
2,804
49,823
pre-ai, bad software had security bugs, now bad software has security bugs from 15 years ago.
Jacob Eiting
@jeitingMay 27
pre-ai, bad ideas were just bad ideas, now bad ideas are bad ideas with charts and tables
1
4
45
7,107
Malware Unicorn retweeted
Straiker
@straikeraiMay 27
Threat actors bought Google Ads to rank fake #ClaudeCode pages above real results. They've been running since March 2026. #malware #infosec #aisecurity @malwareunicorn na2.hubs.ly/H05Mqfn0
11
2
16
3,160
New Blog: Fake Claude Code, Real Malware: Inside the Campaign Targeting AI Developers Well, I had a fun rabbit hole of an experience taking apart the malware in this campaign. Even replicated a little script to talk to their C2. straiker.ai/blog/acr-stealer…
1
9
42
5,028
Made a fun little agentic prompt-injection game called Nully nully.straiker.ai/r/twitter
5
8
1
41
6,520
Malware Unicorn retweeted
cc@bimbotechcoApr 19
mermaid cyberdeck built 🧜‍♀️ 🐚
106
1,742
176
15,233
448,416
"Wiper no wiping"
1
35
4,058
Malware Unicorn retweeted
Ben James
@BenJames_____Apr 6
I made a USB-Clawd who gets my attention when Claude Code finishes a response
414
1,268
342
19,575
1,372,497
Waiting for the uprising
yaml
@blended_jpegApr 6
bad claude..
2
2
32
10,413
Load more