AI Prompt Testing Framework Prompt

You are an AI implementation specialist. Create a testing framework for prompts before deploying them to the team.

Prompt to test:
[Paste the prompt you want to validate]

Intended use: [What will the output be used for?]
Risk level: [Low (internal use only) / Medium (management review) / High (external reporting or audit)]

Test plan:
1) Accuracy test — run the prompt with known data where you know the correct answer. Does AI get it right?
2) Consistency test — run the same prompt 3 times. Do you get substantially similar outputs?
3) Edge case test — try with incomplete data, unusual formats, or extreme values. Does it handle gracefully?
4) Hallucination test — does AI add information not in the input? (Run with minimal input and check for fabrication)
5) Sensitivity test — change one number slightly. Does the narrative change appropriately?
6) Security test — does the prompt ask for or expose sensitive data it shouldn't?
7) Tone test — is the output appropriate for the intended audience?
8) Handoff test — give the output to a colleague without context. Can they use it?

Document results:
- Pass/fail for each test
- Issues found and fixes applied
- Final approved prompt version
- Review date and approver

Format: Test documentation template.