# Find the edge cases and failure states in my feature spec

A structured list of edge cases, empty and error states, and open questions to settle before engineering starts.

- Topic: Product & UX Design
- Framework: RTF
- Level: Standard
- Tags: Analyze, Plan, Fill-in Template
- Source: https://www.thinkwithprompt.com/prompts/find-the-edge-cases-and-failure-states-in-my-feature-spec

## Prompt

```text
Review my feature spec and find the edge cases and failure states we have not covered.

**Feature spec or user flow:** [PASTE SPEC, STORIES, OR FLOW STEPS]
**Platforms:** [web / iOS / Android / API]
**User types and permissions:** [e.g. admin, member, guest, free vs. paid]
**Integrations or data it depends on:** [SYSTEMS, THIRD PARTIES]

Check systematically across these lenses:
- Data: empty, one item, huge volumes, duplicates, special characters, very long text, deleted or archived records.
- Users and permissions: role changes mid-flow, shared accounts, removed users, plan downgrades.
- Timing: concurrent edits, slow or lost connection, timeouts, retries, double-clicks, time zones, daylight saving.
- States: first-time use, partially completed flows, undo, cancel, back button, session expiry.
- Dependencies: third-party outage, rate limits, failed payments or syncs.
- Migration: existing users and data when this ships.
- Accessibility and localization: keyboard-only, screen readers, right-to-left, text expansion.

Return:
1. A table: Scenario | What happens today per the spec (or "undefined") | Recommended behavior | Severity (high / medium / low).
2. The 5 open questions I must answer before development, in priority order.
3. Acceptance criteria to add for the high-severity cases, in Given/When/Then format.

Only flag cases relevant to this feature; skip generic padding.
```

## What to fill in

- PASTE SPEC, STORIES, OR FLOW STEPS
- web / iOS / Android / API
- e.g. admin, member, guest, free vs. paid
- SYSTEMS, THIRD PARTIES

Free to copy and use. From ThinkWithPrompt (https://www.thinkwithprompt.com/), a library of 2,663 curated AI prompts.