Prompt Library
Agents

Agent Eval Set Builder

Test cases that catch regressions before users do.

The Prompt
My agent does: [describe]. Its tools: [list].
Build an eval set:
1. 10 representative tasks across difficulty levels
2. 5 adversarial cases (ambiguous, missing info, prompt injection, out-of-scope)
3. For each: the pass criteria — behavior, not exact wording
4. Which 3 to run on every deploy vs. nightly
Fill in before using
[describe][list]

How to use it

  • → Paste the prompt into your model of choice — it works with any frontier model.
  • → Replace every bracketed placeholder with your specifics. Concrete inputs beat vague ones every time.
  • → If the model asks a clarifying question, answer it — these prompts are designed to invite that.
  • → Iterate on the output, not the prompt. The structure is already tuned.
evalsQA

More Agents prompts