DEV Community

Humanbound

The trust layer for what AI does next

Run a free security scan on your Agent Organization Settings Admin

We test every AI agent in your environment, custom-built or vendor-embedded. Open-source and CLI-first: your engineers can run a full assessment before your next deploy, not after your next incident.

Attacker agents now turn fresh patches into exploits on their own.

Attacker agents now turn fresh patches into exploits on their own.

6
Comments
4 min read
NEAR AI was breached through the engine running its agents, not the prompt

NEAR AI was breached through the engine running its agents, not the prompt

2
Comments
4 min read
A jailbroken AI agent spent 14 days inside a telecom, and the exploits were all old

A jailbroken AI agent spent 14 days inside a telecom, and the exploits were all old

2
Comments
3 min read
Not every model makes a good attacker: OpenAI vs DeepSeek for red-teaming

Not every model makes a good attacker: OpenAI vs DeepSeek for red-teaming

Comments
5 min read
Prompt adware, poisoned AGENTS.md files and the config layer nobody is guarding

Prompt adware, poisoned AGENTS.md files and the config layer nobody is guarding

1
Comments
3 min read
Your red-team model shouldn't grade its own homework

Your red-team model shouldn't grade its own homework

2
Comments
6 min read
A tampered Claude Code extension, five broken coding-agent browsers, and agents nobody owns

A tampered Claude Code extension, five broken coding-agent browsers, and agents nobody owns

Comments
4 min read
LiteLLM getting hijacked again and the AI gateway is the prize

LiteLLM getting hijacked again and the AI gateway is the prize

2
Comments
3 min read
loading...