Map software edge cases with AI
Describe how a feature should behave at its limits, how users recover when something fails and which scenarios deserve the earliest testing.
When you map software edge cases with AI, a small feature makes a good starting point. Consider a file upload that appears to have failed just as the connection drops. The deliver-edge-cases skill helps a product team specify what the user sees, whether the file may already exist and what a retry should do.
The result is a feature-level catalogue. It gives designers the missing states to draw, engineers the expected behaviour to discuss and testers a prioritised set of scenarios. Assumptions about limits can be checked against the actual product before the team commits to them.
What it does
- Identify boundary values on either side of an agreed limit.
- Distinguish validation problems from timeouts and interrupted requests.
- Describe what should happen when two updates arrive together.
- Specify a recovery path that preserves work wherever possible.
- Rank a scenario by its likely impact as well as its frequency.
Install and use
- Install the complete skill folder in your agent. No external service or test runner is required. Provide the feature scope, normal flow and known constraints.
- Ask, “Map the edge cases for changing a delivery slot, including the last available slot being taken while the customer is still choosing.”
- Review the catalogue with design and QA. Settle open product decisions before turning scenarios into tests.
FAQ
Source
Made by Jonathan Prisant (Product on Purpose). Original skill: deliver-edge-cases, published on GitHub under the Apache-2.0 licence.
View on GitHub