Short answerDecide in advance: a pass rate on your test set, a list of must-pass questions, working tools, and a named person who signs off. Without criteria "done" is an opinion.
Define before you start
- Test set: 20 real questions (prompt library).
- Pass threshold: for example, at least 90 percent judged acceptable (pick your own number).
- Must-pass items: refunds, pricing, safety and refusal questions must be 100 percent correct.
- Tone check: reviewed by the owner.
- Tools: each connected action works end to end.
- Access: intended users can reach it.
Sign-off
One named person records the results and approves: tracker. If criteria fail, fix or roll back.
Scoring method: testing guide.
Check your GPT before you move it
Paste your instructions, file names and Actions schema into the free checker. It runs in your browser and builds a portable bundle. Nothing is uploaded.
Open the free checker