Which coding agents are supported?

Claude Code, Cursor, and Codex through the local broker — plus anything MCP-compatible on the same governed endpoint.

Do developers paste API keys anywhere?

No. Sign-in is a device-authorization flow; credentials stay in the vault and are injected server-side at call time.

Can we cut an agent off quickly?

Yes — disable it for the organization and the next tool call fails closed. Live sessions stay visible in the console throughout.

Does this slow developers down?

No — setup is a copy-paste config and one sign-in. After that the agent works exactly as before, just through a governed endpoint instead of raw keys.

Can we see who is running agents right now?

Yes — live broker sessions are visible in the console as they happen, each one attributed to the developer who started it.

Use case · Coding Agents

Govern coding agents on developer machines

Claude Code, Cursor, and Codex work with real repos and systems — give them brokered, revocable access instead of raw keys.

Get a demo

The problem

Coding agents read repos, call APIs, and ship changes — powered by whatever credentials live on the laptop.

Raw keys pasted from ~/.env and dotfiles

No inventory of agents or what they reach

Production touched from personal machines

No switch to flip when something goes wrong

How a laptop agent gets brokered

Enable

You switch the agent on

Claude Code, Cursor, and Codex are enabled per organization — nothing connects uninvited.

The developer authenticates

A device flow ties the session to a real person. No keys pasted anywhere.

Broker

Calls come through Fabriq

Credentials inject server-side at call time; dotfiles stay empty.

Revoke

One switch cuts it off

Disable the agent org-wide and its next call fails closed.

See it

💻 DEVELOPER LAPTOP

Claude Code raw keys · reads all

📁 ~/code/appreadable

🔑 ~/.sshreadable

📄 ~/Documentsreadable

AGENTIC FABRIQ

MCP broker · file & tool scopes

GitHub

prod API

Slack

Capabilities

What you get

Admins enable each coding agent.

Claude Code, Cursor, and Codex are switched on per org — nothing connects uninvited.

A broker instead of raw keys.

Copy-paste config; credentials stay in the vault, never in dotfiles.

Sessions visible live.

Every active broker session shows in the console, attributed to its developer.

A kill switch that actually kills.

Disable an agent org-wide and its next call fails closed — no key rotation scramble.

Every call identified to the developer.

Same permissions and the same audit trail as the rest of the fleet.

The whole tool hub, on the laptop.

Whatever the org has connected — repos, Slack, internal APIs — reachable through one endpoint.

The payoff

The agent on the laptop is still your agent — broker it, watch it, revoke it.

Get a demo

Questions

Govern coding agents on developer machines

Coding agents read repos, call APIs, and ship changes — powered by whatever credentials live on the laptop.

What you get

Common questions