Testing

Test what you build. Write test plans, run them, and file what they turn up as bugs on the project the work belongs to.

What it is
  • Test plans for your products, written where the products live
  • Runs that record what passed, what failed, and when
  • Bugs — issues on the project the work belongs to
Why it exists

Build with agents

Everything your AI agents need to become real software.

Agents write code faster than anyone can check it by hand. The answer is not to read more diffs — it is to say up front what “working” means, and to keep the record of whether it still does.

What it does

Testing is where a product’s quality is written down: the plans that define what must hold, the runs that check it, and the bugs that come out the other side.

  • Write test plans against a product, and version them with it.
  • Record runs, so a regression has a date and a diff.
  • A bug is an issue on a project — it files where the fix will happen.
How to use it
  1. Pick a product and write the plan that says what must hold.
  2. Run it — by hand, or hand it to an agent.
  3. File what failed as a bug; it lands on the project as an issue.