These cases test what an application must verify when a model produces code, a query, an image, a transcript, a video summary, or a current claim. Choose the response that preserves the evidence boundary.
Review the applications
- Coding prompts: name the files, behavior, and proof
- SQL prompts: generate a query under database-enforced limits
- Image prompts: specify the visual contract and inspect the pixels
- Audio prompts: preserve speakers, timestamps, and uncertain words
- Video prompts: cite the moment and the observation
- Research prompts: bind claims to a date and evidence record
Common Mistakes
- Do not accept a generated artifact from its prompt alone.
- Do not execute a query or tool effect based only on model text.
- Do not turn uncertain media evidence into a confirmed cause.