Skip to content

7 · Save time with templates & GitHub

What you'll do

Two shortcuts for jobs you'd otherwise type out by hand: save a template for work you do often, and import issues straight from GitHub.

Templates: save a job to reuse

If you keep typing the same kind of job — "add a test for…", "update the docs for…" — save it once as a template and reuse it.

Save one

On the Work screen, after you've written a job you'll want again, save it as a template and give it a name.

Use one

When creating a job, pick your saved template. It fills in the job for you; adjust the details and start it. Templates can leave blanks for the parts that change each time, so the same shape fits many jobs.

Picking a saved template when creating a job.

Keep your template list short and sharp. A template you have to heavily rewrite every time isn't saving you anything — delete it.

Import from GitHub

If your to-do list already lives as GitHub issues, you don't need to retype them. TofuFactory can import an issue as a job — the issue's title and description become the job, so the AI works from the same write-up your team already has.

Importing a GitHub issue as a job.

A nice rhythm: decide in GitHub which issues are ready, import the ready ones, and let the factory take them through the normal check-and-merge process.

Putting it together

  1. In GitHub, mark an issue as ready.
  2. Import it into TofuFactory.
  3. Apply a template (for example, one that includes your usual "must include a test" note).
  4. Start it, and let the safety check confirm the result.

You should now see

  • A saved template that fills in a job for you.
  • A GitHub issue pulled in as a job without retyping it.
  • A quicker path from "I know what I need" to "it's running."

If something's not right

Problem What to do
A template came out with blanks Fill in its blanks when you pick it; an empty blank stays empty.
GitHub import didn't work Check the issue link and that TofuFactory has access — a private repository needs permission.
The imported job is vague The AI only gets the issue's text. A thin issue makes a thin job — add detail, or apply a template that spells out what "done" means.

Next

→ 8 · Chain jobs into pipelines — when one job should lead into the next.