Prompt with purpose, context, and review

Purpose: Turn safe notes into a draft follow-up.

Expected output: Decisions, action items, owners, dates, open questions, and uncertain points.

Iteration tip: Ask the model to separate confirmed items from items needing confirmation.

Verification step: Check names, deadlines, decisions, and commitments before sending.

Privacy warning: Do not include confidential, personnel, legal, medical, financial, security, customer, or child-related notes.

Unsuitable use: Do not use for sensitive meetings or official records without approved process.

Using these safe meeting notes, draft a follow-up for [audience]. Include confirmed decisions, action items, owners, due dates, open questions, and items needing confirmation. Do not invent decisions or commitments. Mark uncertainty clearly.

Purpose: Draft a low-risk email from facts the sender has approved.

Expected output: A concise draft with a clear next step.

Iteration tip: Ask for a warmer or more direct tone after checking facts.

Verification step: Confirm policies, promises, names, dates, and attachments.

Privacy warning: Do not include private customer, employee, client, medical, legal, or financial details.

Unsuitable use: Do not use to send legal notices, HR actions, medical advice, financial commitments, or crisis messages.

Draft a [tone] email for [audience] about [topic]. Use only these approved facts: [approved facts]. Include a clear next step. Do not invent promises, policies, prices, deadlines, or reasons. List anything I should verify before sending.

Purpose: Find bottlenecks and review points in a low-risk workflow.

Expected output: Steps, bottlenecks, risks, possible AI support, and human checkpoints.

Iteration tip: Ask the model to rank suggestions by risk before usefulness.

Verification step: Confirm the workflow with people who do the work.

Privacy warning: Use fictional or approved process notes only.

Unsuitable use: Do not use for regulated, safety-critical, HR, finance, health, legal, or security workflows without qualified review.

Review this low-risk workflow: [workflow name]. Steps: [safe steps]. Pain points: [pain points]. Identify bottlenecks, handoffs, risks, possible AI assistance, required human checkpoints, and success measures. Do not recommend full automation unless risks are low and review is clear.

Purpose: Create a draft checklist for a low-risk recurring decision.

Expected output: Criteria, required information, red flags, and escalation triggers.

Iteration tip: Ask what could go wrong if the checklist is used too mechanically.

Verification step: Have the decision owner review criteria and escalation points.

Privacy warning: Do not include personal records or sensitive case details.

Unsuitable use: Do not use for rights, benefits, employment, health, legal, financial, education-discipline, or safety decisions without official policy and qualified review.

Help draft a checklist for a low-risk decision: [decision]. Audience: [users]. Include required information, criteria, red flags, escalation triggers, and what not to decide by checklist. Mark assumptions and items needing policy review.

Purpose: Explain a topic in plain language.

Expected output: Short explanation, example, key terms, and check questions.

Iteration tip: Ask for a simpler version or a different example.

Verification step: Compare facts with a trusted source.

Privacy warning: Use public topics, not private situations.

Unsuitable use: Do not use as a substitute for teacher instructions or professional advice.

Explain [topic] to [audience level] in clear language. Include one everyday example, three key terms with definitions, two common misunderstandings, and five self-check questions. Mark any facts I should verify.

Purpose: Practice recall and understanding.

Expected output: Questions, answer key, explanations, and weak-area suggestions.

Iteration tip: Ask for easier or harder questions after answering.

Verification step: Check the answer key against course material or trusted sources.

Privacy warning: Do not paste private student records or graded work.

Unsuitable use: Do not use to cheat on tests or assignments.

Create a practice quiz on [topic] for [level]. Include [number] questions using a mix of multiple choice, short answer, and explanation questions. Provide an answer key with brief explanations and list what I should review if I miss each question.

Purpose: Plan a short learning sprint.

Expected output: Weekly goals, practice tasks, evidence artifact, and reflection questions.

Iteration tip: Ask to reduce the time requirement if the plan is too large.

Verification step: Compare resource suggestions with official or trusted sources.

Privacy warning: Do not include sensitive personal details.

Unsuitable use: Do not use for licensing, medical, legal, financial, or urgent decisions without qualified guidance.

Create a [number]-week learning plan for [skill] for a learner who can spend [hours per week]. Use public or low-risk practice tasks. Include weekly goals, exercises, one evidence artifact, reflection questions, and safety checks. Do not promise outcomes.

Purpose: Find likely misunderstandings about a topic.

Expected output: Misconceptions, corrections, examples, and study prompts.

Iteration tip: Ask for misconceptions specific to beginners or a certain role.

Verification step: Check corrections against course material or trusted references.

Privacy warning: Use general learning needs, not private student details.

Unsuitable use: Do not use to diagnose a learner or make education-placement decisions.

List common misconceptions about [topic] for [audience]. For each one, explain the correction in plain language, give a safe example, and suggest one practice question. Mark any claims that need verification.

Purpose: Make a draft clearer and easier to read.

Expected output: Revised draft plus a list of changes.

Iteration tip: Ask for a version for a different audience.

Verification step: Check that meaning did not change.

Privacy warning: Do not paste confidential drafts into unapproved tools.

Unsuitable use: Do not use to remove required warnings, terms, citations, or policy details.

Rewrite this safe draft for [audience] at about an eighth-grade reading level. Keep the meaning, facts, and required details. Improve structure and clarity. After the rewrite, list what you changed and what I should verify.

Purpose: Check whether a message sounds respectful and clear.

Expected output: Tone risks, suggested edits, and revised draft.

Iteration tip: Ask for more direct, warmer, or more neutral options.

Verification step: Confirm the revised tone fits the real relationship.

Privacy warning: Replace private names and details before using.

Unsuitable use: Do not use to manipulate, impersonate, or hide responsibility.

Review this safe message for tone, clarity, and possible misunderstandings. Audience: [audience]. Purpose: [purpose]. Suggest edits that keep my meaning and voice. Do not add facts or promises. List any sentences that could sound harsh, vague, or too certain.

Purpose: Organize a draft before writing.

Expected output: Structured outline, missing questions, and source needs.

Iteration tip: Ask for a shorter outline if it is too complex.

Verification step: Check that each section has real support.

Privacy warning: Use only public or approved topic information.

Unsuitable use: Do not use to generate fake citations or unsupported arguments.

Create an outline for [document type] about [topic] for [audience]. Include purpose, section headings, key points, examples, source needs, and questions I must answer before drafting. Do not invent facts or citations.

Purpose: Check a public-facing draft for accessibility and usability issues.

Expected output: Accessibility concerns and improvement suggestions.

Iteration tip: Ask for a version optimized for scanning or screen-reader clarity.

Verification step: Review against WCAG and human testing when needed.

Privacy warning: Do not include private user data.

Unsuitable use: Do not treat this as a formal accessibility audit or legal compliance opinion.

Review this safe public-facing draft for accessibility and usability. Check plain language, headings, link text, image descriptions needed, table clarity, reading order, and confusing instructions. Suggest improvements and list what should be manually tested.

Purpose: Break a claim into parts that can be checked.

Expected output: Claim parts, source needs, assumptions, and uncertainty.

Iteration tip: Ask the model to rank claims by risk or importance.

Verification step: Check claims with trusted sources outside the model.

Privacy warning: Use public claims, not private allegations or sensitive records.

Unsuitable use: Do not use as the only source for legal, medical, tax, financial, or safety claims.

Break this claim into checkable parts: [claim]. For each part, say what kind of source would be needed, what assumptions are involved, what could be uncertain, and what would change the conclusion. Do not decide truth without evidence.

Purpose: Evaluate whether sources fit a question.

Expected output: Source roles, strengths, limits, and gaps.

Iteration tip: Ask what source types are missing.

Verification step: Open and read the sources yourself.

Privacy warning: Do not include private documents unless the tool is approved.

Unsuitable use: Do not use to launder weak sources into strong claims.

Review these source descriptions for the question: [question]. Sources: [source titles and short notes]. For each source, identify source type, relevance, currency, possible bias, limits, and what claim it can support. Do not invent source content.

Purpose: Draft a short brief from source notes.

Expected output: Summary, facts, uncertainty, source map, and next questions.

Iteration tip: Ask for a version with stronger caveats.

Verification step: Compare every factual sentence with the source notes.

Privacy warning: Use public source notes only.

Unsuitable use: Do not use for professional advice or high-stakes decisions.

Using only these source notes, draft a brief for [audience] about [question]. Include a short summary, key facts, source mapping, uncertainty, conflicting evidence if present, and next questions. Do not add facts not in the notes.

Purpose: Slow down before sharing a news claim.

Expected output: Check questions, source types, and sharing caution.

Iteration tip: Ask for what would be needed to verify the claim today.

Verification step: Check current reputable sources before sharing.

Privacy warning: Do not include private messages or allegations about private people.

Unsuitable use: Do not use to confirm breaking news without current source checking.

Help me sanity-check this public news claim before sharing: [claim]. Break it into checkable parts, list source types needed, identify emotional or manipulative wording, and write a cautious version that does not overstate what is known.

Purpose: Organize real skills and gaps.

Expected output: Proven skills, developing skills, examples, artifacts, and learning gaps.

Iteration tip: Ask for a shorter version focused on one role.

Verification step: Compare with real job postings and honest evidence.

Privacy warning: Use a non-sensitive summary, not a full private resume.

Unsuitable use: Do not use to create fake experience, fake credentials, or hidden keyword stuffing.

Using this non-sensitive experience summary, organize a truthful skill inventory for [target role]. Separate proven skills, developing skills, examples I can explain, possible evidence artifacts, and gaps to learn. Do not invent experience, tools, credentials, metrics, employers, or outcomes.

Purpose: Plan a low-risk project.

Expected output: Goal, tasks, timeline, risks, checkpoints, and review plan.

Iteration tip: Ask to reduce scope if the plan is too large.

Verification step: Confirm deadlines, dependencies, and owner responsibilities with real people.

Privacy warning: Do not include confidential project details.

Unsuitable use: Do not use for safety-critical, legal, medical, financial, or regulated project decisions without qualified review.

Create a small project plan for [project] with [timeframe]. Include goal, audience, tasks, owners, dependencies, risks, human checkpoints, success criteria, and a review date. Use only the safe information provided and mark assumptions.

Purpose: Plan a short learning sprint.

Expected output: Weekly practice, artifact, reflection, and safety checks.

Iteration tip: Ask for a version for beginner, job seeker, or employed learner.

Verification step: Check resources and credential claims independently.

Privacy warning: Keep personal constraints general.

Unsuitable use: Do not use to make expensive education or career decisions without additional human and official-source review.

Build a [duration]-day learning sprint for [skill]. I can spend [hours per week]. Include weekly practice, one artifact, reflection questions, feedback ideas, and safety checks. Do not promise results. List assumptions I should verify.

Purpose: Decide whether a task is suitable for AI support.

Expected output: Risk level, safer role for AI, checkpoints, and escalation triggers.

Iteration tip: Ask for a more conservative version.

Verification step: Check tool policy, data rules, and domain rules before acting.

Privacy warning: Describe the task generally; do not include sensitive details.

Unsuitable use: Do not use to approve high-stakes AI use by yourself.

Review this proposed AI use: [general task description]. Assess stakes, reversibility, sensitive data, error detectability, judgment required, affected people, and policy questions. Recommend whether AI should assist, automate with review, stay human-led, or not be used. List escalation triggers.

Purpose: Prepare for a respectful conversation without scripting deception.

Expected output: Goals, listening questions, points to clarify, and tone suggestions.

Iteration tip: Ask for a calmer or shorter version.

Verification step: Make sure the final words are yours and fit the relationship.

Privacy warning: Remove names and private details.

Unsuitable use: Do not use to manipulate, threaten, diagnose, or avoid urgent professional help.

Help me prepare for a respectful conversation about [topic] with [audience]. My goal is [goal]. Suggest opening points, listening questions, possible misunderstandings, and a calm closing. Do not script manipulation, threats, or false claims. Mark what I should decide myself.

Purpose: Organize options for a low-risk decision.

Expected output: Options, tradeoffs, missing information, and next questions.

Iteration tip: Ask to add a no-action option.

Verification step: Confirm facts and consult qualified help for serious decisions.

Privacy warning: Keep details general and non-sensitive.

Unsuitable use: Do not use for legal, medical, tax, financial, safety, or emergency decisions.

Help map options for this low-risk decision: [decision]. Include options, benefits, risks, unknowns, people affected, information needed, and a next-step checklist. Do not make the decision for me. Mark any area needing qualified advice.