📚 AI Product Management, Evals & Prompt Systems

Use this skill to turn messy product context into a useful synthetic data generation plan with clear reasoning and next steps.

Use this skill to review responsible-ai review checklist and turn the findings into specific risks, tradeoffs, and next actions.

Use this skill to design a red-team session to find adversarial inputs that break an AI feature.

Use this skill to turn messy product context into a useful rag quality evaluation plan with clear reasoning and next steps.

Use this skill to review prompt-injection threat review and turn the findings into specific risks, tradeoffs, and next actions.

Use this skill to turn messy product context into a useful prompt-feature specification with clear reasoning and next steps.

Use this skill to documents a production prompt with role, tone, formatting, refusals, and version notes.

Use this skill to builds a 90-minute workshop teaching a team practical prompt engineering.

Use this skill to design an A/B test for prompt variants with sample sizes and eval pre-registration.

Use this skill to build a useful model-selection trade-off matrix with explicit inputs, assumptions, scoring logic, and decision criteria.

Use this skill to picks the right model for a feature with cost, latency, quality, and switching plan.

Use this skill to build a useful model drift monitoring plan with explicit inputs, assumptions, scoring logic, and decision criteria.

Use this prompt to turn messy product context into a useful llm-as-judge eval setup with clear reasoning and next steps.

Use this skill to turn messy product context into a useful llm cost vs quality analysis with clear reasoning and next steps.

Use this prompt to turn messy product context into a useful human-in-the-loop flow design with clear reasoning and next steps.

Use this skill to turn messy product context into a useful hallucination mitigation strategy with clear reasoning and next steps.

Use this skill to turn messy product context into a useful golden test-set builder with clear reasoning and next steps.

Use this skill to draft a clear fine-tune vs prompt vs rag memo that frames the decision, evidence, tradeoffs, and recommendation.

Use this skill to decides whether to upgrade to a new model version by comparing eval performance.

Use this skill to create an eval specification with tasks, graders, datasets, and pass criteria for an AI feature.

Use this skill to turn messy product context into a useful eval set & rubric designer with clear reasoning and next steps.

Use this prompt to curates a balanced eval dataset from production logs without leaking sensitive data.

Use this skill to audits an AI feature's cost-per-query and recommends optimization tactics.

Use this skill to turn messy product context into a useful confidence & uncertainty ux spec with clear reasoning and next steps.

Use this skill to review bias & fairness audit prompt and turn the findings into specific risks, tradeoffs, and next actions.