# UVP Free AI Prompt Library v2

Unshaken Voice Productions

**49 public prompts and samples**

This edition contains 25 Basic prompts, 18 Advanced prompts, one Expert build sample, and five Image Generation samples. It is the free layer of a larger UVP prompt-systems catalog.

## Use standard

1. Replace bracketed fields with information you actually know.
2. Provide source material before asking for source-dependent conclusions.
3. Tell the model to mark assumptions rather than invent missing facts.
4. Review output before using it with customers, employees, students, production systems, or the public.
5. Do not paste credentials, restricted personal records, private client data, or unauthorized material into an AI system.
6. For tool-enabled agents, begin read-only and require approval before destructive, paid, credentialed, production, or externally visible actions.

## 25 Basic Prompts

1. **Clarify a Goal** — Turn this goal into a clear outcome, constraints, success criteria, and the next three actions: [GOAL]. Mark missing information instead of inventing it.
2. **Summarize for a Busy Reader** — Summarize [TEXT] in five bullets, then list decisions, risks, assumptions, and open questions.
3. **Make a Checklist** — Turn [PROCESS] into a checklist with owner, required input, completion test, and failure note for each step.
4. **Compare Options** — Compare [OPTIONS] by cost, effort, risk, reversibility, time-to-value, and likely benefit. State what is missing before recommending.
5. **Draft a Simple SOP** — Draft an SOP for [TASK] with purpose, scope, inputs, ordered steps, review/approval points, handoff, and quality check.
6. **Prepare Meeting Notes** — Organize [NOTES] into decisions, actions, owners, deadlines, blockers, follow-up questions, and approval items.
7. **Review for Accessibility** — Review [CONTENT] for plain language, headings, keyboard usability, contrast concerns, alt-text needs, mobile readability, and complexity.
8. **Find Missing Questions** — Before we build/publish/buy/commit to [IDEA], identify material questions by user need, product value, technical risk, legal risk, operations, and cost.
9. **Create a Research Plan** — Create a research plan for [TOPIC] with source types, search terms, verification, comparison criteria, and final deliverable.
10. **Explain for a Beginner** — Explain [CONCEPT] in plain language with one example, one non-example, three common mistakes, and three checks for understanding.
11. **Turn Feedback into Work** — Group [FEEDBACK] into bugs, content, design, features, risks, and questions, then rank by urgency and effort.
12. **Plan a Small Launch** — Create a launch checklist for [PRODUCT/SERVICE] with must-have, should-have, later, go/no-go, and rollback actions.
13. **Draft a Customer Email** — Draft a short respectful email about [SITUATION] using only [FACTS], no overpromises, and one clear next step.
14. **Create a Content Outline** — Outline useful content about [TOPIC] with audience, promise, sections, examples, proof points, and limitation.
15. **Check a Workflow** — Review [WORKFLOW] for unclear ownership, approvals, privacy risk, handoff risk, failure points, and unnecessary manual work.
16. **Make Assumptions Visible** — List assumptions in [PROPOSAL], mark confirmed/likely/uncertain/risky/unsafe, and name the five to verify first.
17. **Prepare a Handoff** — Create a handoff brief for [PROJECT] with context, state, files/links, decisions, questions, risks, and exact next action.
18. **Test a User Path** — Write a manual test plan for [PATH] covering happy, empty, error, mobile, accessibility, and rollback states.
19. **Improve a Service Description** — Rewrite [DRAFT] to clarify who it helps, inclusions, non-promises, proof, and next step.
20. **Reflect Before Automating** — For [TASK], identify human decision, required data, failure cost, approval point, safest pilot, and what stays human.
21. **Extract Requirements** — Extract explicit requirements, goals, constraints, assumptions, unresolved questions, and out-of-scope items from [MATERIAL].
22. **Write a Decision Memo** — Create a memo for [DECISION] with options, evidence, assumptions, tradeoffs, recommendation, reversible next step, and change conditions.
23. **Build a Risk Register** — For [PROJECT], list risks with trigger, impact, likelihood evidence, mitigation, owner, and verification.
24. **Turn an Idea into User Stories** — Convert [IDEA] into user stories grouped by role, tied to observable needs and one acceptance check each.
25. **Source Claim Check** — Label each claim SUPPORTED, UNSUPPORTED, CONFLICTING, or NEEDS VERIFICATION based only on [SOURCES].

## 18 Advanced Prompts

1. **Root-Cause Debugger** — Diagnose a failing test/error from evidence, propose the smallest safe fix, list regression risk, and provide verification.
2. **Clean Refactor Pass** — Refactor for readability/testability without behavior change; explain changes, uncertainty, regression risk, and tests.
3. **Production REST API Contract** — Define endpoints, methods, payloads, validation, auth assumptions, errors, rate limits, and examples.
4. **Senior Pull Request Review** — Review correctness, edge cases, security, data handling, accessibility, performance, maintainability, and tests.
5. **Complete Test Suite Builder** — Design happy, edge, invalid, failure, and regression tests; identify branches blocked by design.
6. **Normalized Database Schema** — Design tables, keys, indexes, constraints, migrations, seeds, and future migration considerations.
7. **Performance Diagnosis Pass** — Separate algorithmic, database, network, render, cache, and runtime bottlenecks; profile before fixing.
8. **Endpoint Hardening Review** — Check authn/authz, validation, injection, rate limiting, uploads, secrets/logs, error leakage, API abuse, and replay.
9. **Production CI Pipeline Planner** — Define install/cache, lint, type checks, tests, build, security, previews, blockers, rollback, and hotfix path.
10. **Safe System Migration Plan** — Stage migration with backup, compatibility, rollback, validation, customer impact, training, and cutover criteria.
11. **System Architecture Mapper** — Produce an observed current-state map before judging design quality.
12. **Dependency and Tool Audit** — Separate required/optional dependencies, unsupported items, licensing, install risks, prerequisites, and fallbacks.
13. **AI-to-AI Build Handoff** — Create canonical sources, dependency-safe tasks, context capsules, guardrails, verification matrix, rollback, and progress ledger.
14. **Project Explanation Builder** — Turn a rough idea into purpose, audience, capabilities, exclusions, standards, sources, approvals, success, and blockers.
15. **Master Prompt Audit** — Find ambiguity, missing information, conflicts, assumptions, weak controls, safety/privacy gaps, unclear deliverables, and premature execution.
16. **Deliverable Compliance Audit** — Check every requirement against evidence and list partial/missing work, unverified claims, assumptions, privacy, and approvals.
17. **Context Compaction Handoff** — Preserve canonical decisions, state, paths, risks, sources, unresolved issues, and next action while dropping superseded chatter.
18. **Release Readiness Review** — Check whether a release is safe to ship from scope, tests, dependencies, migrations, observability, security, rollback proof, and blockers.

## Expert Free Sample — Governed MVP Builder

```text
You are a governed software build agent. Build a bounded, original MVP for [PRODUCT] from the supplied requirements.

Start by inspecting the actual environment, repository, available tools, dependencies, and constraints. Establish the smallest demonstrable vertical slice before expanding scope. Preserve user data and existing working behavior.

Ask approval before paid services, credentials, destructive changes, production deployment, external publishing, or irreversible migration.

Implement modularly, add tests for the core path, run available type/lint/build/test checks, document exact install/run commands, record files changed, state what could not be verified, and finish with: result status, project path, stack, features implemented, validation performed, known limitations, rollback/recovery notes, and the next bounded build pass.

Do not claim completion when tests or required outputs have not been verified.
```

## Five Image Generation Samples

### 1. Extraordinary Logo + Identity System
Create an original premium identity system for [BRAND] serving [AUDIENCE]. Develop one distinctive visual idea. Show primary logo, monochrome mark, icon-only mark, horizontal lockup, and small-size preview. Prioritize clarity, geometry, negative space, typography, and legibility. Avoid trademark imitation, generic clipart, crowded mockups, and decorative effects without purpose.

### 2. Premium Product Hero Advertisement
Create a premium 4:5 hero advertisement from the uploaded product image. Preserve shape, packaging, branding, logos, colors, labels, proportions, materials, artwork, and texture exactly. Use supplied claims only. Avoid altered branding, distorted labels, invented benefits, clutter, and excessive effects.

### 3. Software Program Hero Visual
Create a premium 16:9 software hero for [SOFTWARE] serving [TARGET USER]. Show believable navigation, panels, workflows, search, settings, and realistic data states. Avoid gibberish-heavy UI, fake metrics, random icons, and unsupported features.

### 4. AI Feature Visualization
Visualize [FEATURE] as an understandable flow from input to context to governed action to output. Keep the metaphor restrained and functional. Avoid robots, glowing brains, binary wallpaper, surveillance imagery, and unsupported product claims.

### 5. Premium Infographic / Explainer Visual
Create an editorial infographic about [TOPIC] with [NUMBER] sections, clear hierarchy, concise copy, readable typography, clean diagrams, and accurate labels. Use placeholders instead of inventing missing statistics.

## Where the paid library goes deeper

Premium UVP packages add domain-specific operating rules, routers, evidence/provenance standards, approval levels, dependencies and fallbacks, durable artifact contracts, verification, negative tests, release gates, and implementation handoffs.

Current premium families include Prompt Engineering, Evidence-Driven AI Builder, Advanced Developer Systems, Local Agent Operations, Forge Engineering Build Systems, Visual Generation, Business Operations, Research/Education, Content/Publishing, Context/Continuity, and reusable AI skills.

Visit **unshakenvoice.tech/prompt-library** for current release status and pricing.