TRUESEAL LIVE: Run a deepfake scan yourself

Guardian Agent Blackbox: Keep your AI agents safe from being tricked or compromised.

Join the network with one command.

$curl -fsSL blackbox.umanitek.ai | bash

Why do your agents need protection?

AI agents can run commands, access files, and install software. That power makes them useful. It also makes them targets.

They can get tricked

Hidden 'prompt injection' text buried in a website or file can hijack your agent and change what it does.

They can leak secrets

Attacked agent might share your SSH keys, passwords, or API tokens.

They can run bad code

One wrong command like rm -rf ~/ or a malicious package can takeover or delete your entire system.

Guardian Agent Blackbox.Your agent's real-time defenses.

Prompt injection

Hidden instructions buried in pages, files, and tool output.

Vulnerable dependencies

Packages with known gaps or malicious versions.

Dangerous commands

Shell actions that delete, exfiltrate, or alter systems.

Sensitive file access

SSH keys, tokens, .env, credentials.

Secret exposure

Real keys handled or shipped off-machine.

Suspicious skills

Newly installed skills that misbehave.

Six categories of risks,
handled continuously.

A live dashboard shows every threat as it's caught.

Run and watch findings land in real time - ranked by severity, each traced back to the threat graph that confirmed it.

Guardian AgentTrace live threats dashboard

When one Agent Blackbox finds a threat, every Blackbox in the network gets safer.

Threats that our agent identifies live anonymised on a public, tamper-proof OriginTrail's Decentralised Knowledge Graph, no single party can rewrite. Every Agent Blackbox syncs with the graph. The network hardens together.

Join the Agent Blackbx

Open source. One command install.

One command to protect your agent

It's open-source, runs entirely on your machine, and starts in flag-only mode so nothing is blocked until you say so.

$curl -fsSL blackbox.umanitek.ai | bash
Enable Agent Blackbox and it attaches to every agent on your machine.
Open the live dashboard to watch threats as they're caught.
Prefer to read before you run? Full manual steps are in the README.