How it works

The premise

Coding agents can’t sign in to Firebase apps. Signing up needs an inbox, a password or an OAuth screen, so agents only ever see the logged-out version of your site and can’t check the features they just changed. Making throwaway test accounts clutters your user list, skews analytics and leaks state from one run to the next.

This library fixes that for sites you run. Add it once, and an agent can open a single URL and land in a signed-in sandbox session, using a test identity that stays the same every time and fresh sample data. The data is wiped between runs. The identity sticks around.

What’s included

Named test personas, so an agent comes back as the same user every run

Sample data created for each session and wiped afterward, so runs don’t leak into each other

Separate lanes so several agents can work on one project at the same time

An MCP server and CLI for creating access grants, with the grant endpoint protected by a service key

How it’s built

firebase-agent-sandbox is built on TypeScript, Firebase and MCP.

The repository is 96% TypeScript, 4% JavaScript by volume.

It carries 7 commits since Aug 2026, most recently changed in Aug 2026.

How the work is done

Every project here starts as a written spec, and the spec stays the source of truth once there’s code. The repository keeps a build plan with current status, an append-only decision log of where the code departed from the plan, and a handoff document that describes where things stand, including the unfinished parts.

That’s how one person keeps this many projects moving, and it’s what this page is written from.

Current state

Prototype. Works, but isn’t finished.

The source is private, so the way to check what this page says is to use the product.