Nine ranks. One habit to work on next.

Forge reads your Claude Code sessions and shows where you stand.

Climbing to Example

Initiate

You name the check, your agents prove each change by running it, and you judge the result.

What Forge reads

Reads: Forge reads the Claude Code sessions already on your computer.
Looks up: It looks up your own commits and pull requests through git and gh.
Runs: Forge runs on your computer, where your sessions sit.
Sends: Your results go to your team's Forge. Your prompts go only if you say so.

The rank you hold, and the one you are climbing to.

The next habit to work on, and how to do it.

Every habit links to the talks and writing behind it.

Your engineers are still new to working through agents.

Novice
Your agents write your code in auto mode, from a full brief and your exact build and test commands.
Initiate
You name the check, your agents prove each change by running it, and you judge the result.
Scholar
You spec large work, own the contracts and push back on plans, and your agents research for themselves.

Your engineers' runs finish, get reviewed and land on their own, and the work holds up.

Acolyte
Your agents run until checks pass, get a fresh review and land focused pull requests.
Apprentice
Your corrections become checks and shared skills, your scenarios stay protected, and away runs start from a spec.
Adept
You run many agents at once, split big jobs, compare versions and keep the code uniform.

Real demand starts your engineers' agents, and protected checks decide what ships.

Artisan
Your agents get full environments at scale, scheduled jobs land work, and your slowest step gets automated.
Expert
Areas earn agent review merges through tandem review, and your agents improve your setup from its logs.
Master
Real demand starts your agents, your setup lifts your teammates, and you retest it on each new model.

See each engineer's rank and the habit they are working on.

Example