name: facet-prompt-optimizer description: Use when bootstrapping a facet ground-truth dataset, running initial facet evals, or later optimizing a facet prompt in this repo.
Instructions
Follow the workflow in README.md. Treat the README as the source of truth for commands, defaults, and sequencing.
Agent guardrails
- Do not duplicate workflow instructions here. Update the README first when the workflow changes.
- Use the latest
btCLI and callbt evaldirectly for evals. - Keep customer-owned prompts and generated datasets in local ignored paths such as
.local/facet-optimizer. - Do not treat production facet outputs as ground truth; use generated or reviewed
expectedvalues. - Do not proceed from dataset creation to prompt optimization until the user has reviewed the dataset balance and initial evals.