Hands-on workshop
Govern every AI agent on every laptop
Every developer now runs AI agents: Claude Code, Codex, Claude Desktop. Nobody inventoried them, they run on personal keys and plug into tools nobody approved. In this workshop you take a laptop through the whole story, from “what is running?” to “every call is governed”.
- 10 min hands-on
- 5 chapters
- Your own Linux desktop
- Nothing to install
What you’ll do
- 1
Enroll your laptop · 2 min
A managed laptop: enrolled in acme’s fleet as Alice, with every agent getting its settings from the company.
- 2
See every agent · 1 min
Answer “what AI is running on our laptops?”: every agent, MCP server and skill, tied to a person and a device.
- 3
Agents without keys · 2 min
Use Claude Code and Codex with no API key on the laptop: every call goes through the company’s agentgateway.
- 4
Every call, every cost · 2 min
Find your own calls, with their tokens and cost, by user, model and client, in the gateway’s analytics and logs.
- 5
Same laptop, different rules · 3 min
Sign in as Bob and see the gateway refuse Claude Code while Codex works: policy follows the person, not the laptop.
What you get
A laptop of your own
A Linux desktop in your browser for the session, with Claude Code, Codex and Claude Desktop installed. Nothing to install on yours.
A real fleet
An agentdesktop controller and a Solo Enterprise agentgateway, shared by the room, so you see everyone’s traffic like a security team would.
Ten minutes, guided
Five short chapters: why it matters, what to do, your work checked live, and one quick question. A timer keeps you on track.