01 / Run limits
When should an agent stop?
A failed run can still produce useful work. Stripe, Dropbox, and DoorDash use different limits.
From the cases
Notes
Short observations for people who build agents.
Each note examines a design choice from the catalog. Sources describe what teams report. Our observations explain what those reports may mean for other builders.
02 / Human attention
More comments can mean more work
Uber and HubSpot check review comments before engineers see them.
03 / Agent roles
Separate the search from the check
DoorDash changed how its agents divide a code review. Each design exposed a different problem.
04 / Work state
The worker can stop. The work can continue.
A new worker can continue from a saved record.
05 / Tools and context
Load tools when the task needs them
An agent can find a tool before it loads the details.
06 / Workflow control
Some steps do not need a model
Code can control a step that must follow a fixed rule.
07 / Evaluation
Test the agent on your own work
Past tasks and failures can become repeatable checks.
These notes describe selected cases. They do not establish that one design works best for every team.