Productivity prompts built around real decisions.

Design a reusable prompt system for [RECURRING TASK], used by [USERS]. Typical inputs: [INPUTS]. Required output: [OUTPUT REQUIREMENTS]. Known failures: [KNOWN FAILURES].

Define the task contract: objective, user, source hierarchy, mandatory inputs, optional inputs, missing-information behavior, operating steps, decision rules, boundaries, output schema, examples, and acceptance criteria. Create the master prompt with clearly named variables and instructions that remain portable across ChatGPT, Claude, and Gemini.

Then create a five-case evaluation set covering normal use, incomplete input, conflicting evidence, an edge case, and an unsafe or out-of-scope request. Provide a scoring rubric, version label, change log template, owner, review trigger, and rollback rule. Do not optimize for one polished demonstration.

REVIEWED EXECUTION PROTOCOL
1. Begin by restating the objective, intended audience, supplied evidence, constraints, and missing information.
2. Ask no more than three high-impact clarification questions when missing context would materially change the result. Otherwise continue with clearly labeled assumptions.
3. Work from the supplied facts. Do not invent research, statistics, quotations, customer evidence, rankings, citations, capabilities, or results.
4. Make the reasoning auditable: distinguish evidence, interpretation, recommendation, risk, and open question.
5. Produce a decision-ready deliverable, not generic advice. Prioritize the most valuable actions and explain the trade-offs.
6. Finish with:
   - Assumptions to validate
   - Risks and failure modes
   - Recommended next actions
   - Quality-control checklist

QUALITY STANDARD
The result must be specific enough for an experienced professional to use, concise enough to review, and honest about uncertainty. Reject vague filler, repeated ideas, unsupported superlatives, and fabricated facts.

Design a governed prompt library for [TEAM] covering [USE CASES] across [TOOLS]. Risk levels: [RISK LEVELS]. Current prompts or storage: [CURRENT PROMPTS].

Define the record schema for each prompt: identifier, title, owner, purpose, intended users, model compatibility, required inputs, prompt text, examples, evaluation cases, limitations, approved data classes, status, version, last review, and change history. Propose roles and permissions for authors, reviewers, users, and administrators.

Create submission, test, approval, publication, monitoring, update, deprecation, and incident workflows. Include naming rules, duplicate detection, search tags, adoption metrics, failure reporting, and a 30-day pilot plan. Keep controls proportional to risk and avoid unnecessary bureaucracy.

REVIEWED EXECUTION PROTOCOL
1. Begin by restating the objective, intended audience, supplied evidence, constraints, and missing information.
2. Ask no more than three high-impact clarification questions when missing context would materially change the result. Otherwise continue with clearly labeled assumptions.
3. Work from the supplied facts. Do not invent research, statistics, quotations, customer evidence, rankings, citations, capabilities, or results.
4. Make the reasoning auditable: distinguish evidence, interpretation, recommendation, risk, and open question.
5. Produce a decision-ready deliverable, not generic advice. Prioritize the most valuable actions and explain the trade-offs.
6. Finish with:
   - Assumptions to validate
   - Risks and failure modes
   - Recommended next actions
   - Quality-control checklist

QUALITY STANDARD
The result must be specific enough for an experienced professional to use, concise enough to review, and honest about uncertainty. Reject vague filler, repeated ideas, unsupported superlatives, and fabricated facts.

Create an evaluation suite for this prompt: [PROMPT]. Task: [TASK]. Definition of a good output: [GOOD OUTPUT]. Known failure modes: [FAILURE MODES]. Models or versions to compare: [MODELS].

Design at least twelve test cases covering typical inputs, missing context, noisy input, conflicting sources, boundary conditions, adversarial phrasing, format pressure, and out-of-scope requests. For each case, provide the input, expected behavior, prohibited behavior, critical assertions, and severity if it fails.

Build a scoring rubric for correctness, evidence use, instruction adherence, completeness, usefulness, uncertainty handling, safety, format, latency, and cost where available. Define pass thresholds, reviewer instructions, comparison rules, regression triggers, and how to document a prompt or model update.

REVIEWED EXECUTION PROTOCOL
1. Begin by restating the objective, intended audience, supplied evidence, constraints, and missing information.
2. Ask no more than three high-impact clarification questions when missing context would materially change the result. Otherwise continue with clearly labeled assumptions.
3. Work from the supplied facts. Do not invent research, statistics, quotations, customer evidence, rankings, citations, capabilities, or results.
4. Make the reasoning auditable: distinguish evidence, interpretation, recommendation, risk, and open question.
5. Produce a decision-ready deliverable, not generic advice. Prioritize the most valuable actions and explain the trade-offs.
6. Finish with:
   - Assumptions to validate
   - Risks and failure modes
   - Recommended next actions
   - Quality-control checklist

QUALITY STANDARD
The result must be specific enough for an experienced professional to use, concise enough to review, and honest about uncertainty. Reject vague filler, repeated ideas, unsupported superlatives, and fabricated facts.