skills.rest
Military-Veteran-Team-LPT-Realty
Military-Veteran-Team-LPT-RealtyOFFICIAL
Explore all skills by this author →
GitHub

ai-message-polisher-builder

Builds a full-stack AI message tone rewriting web app with dual variations.

Updated Sep 16, 2026
Search Tags:#prompt engineering#react#web app#trpc#llm prompts#message rewriting#tone generator
One-click install
npx skills add https://github.com/Military-Veteran-Team-LPT-Realty/mvt-manus-public-skills --skill ai-message-polisher-builder-military-veteran-team-lpt-realty
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ai-message-polisher-builder
Source: https://github.com/Military-Veteran-Team-LPT-Realty/mvt-manus-public-skills/tree/main/skills/ai-message-polisher-builder
Command: npx skills add https://github.com/Military-Veteran-Team-LPT-Realty/mvt-manus-public-skills --skill ai-message-polisher-builder-military-veteran-team-lpt-realty

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Building an AI message rewriting tool from scratch involves many scattered decisions: prompt guardrails, tone catalogs, parallel LLM calls, UI layout, branding, testing, and GitHub setup. This Skill encodes the complete end-to-end workflow so the app can be reproduced or adapted without re-deriving every decision. ## Core Features & Use Cases - Server-side LLM architecture: All AI calls run server-side via invokeLLM with a CORE_GUARDRAIL system prompt that forces rewrite-only behavior, plus parallel dual-variation generation using Promise.all. - Tone and style catalog: 12 tone prompts (Friendly, Boss Mode, Legalistic, Gospel Motivation, NLP, Sales, and more) combined with 3 style modifiers (Concise, Elaborate, Formalize) at runtime. - Complete frontend and delivery workflow: React 19 + TypeScript + Tailwind + Express + tRPC stack with a tone grid UI, VariationCard components with word/character counts and copy buttons, a Help/FAQ page, MVT dark navy and gold branding, Vitest tests, and GitHub CI setup. - Use Case: A developer needs to ship a mobile-friendly tool that rewrites a pasted message in a chosen tone and returns two polished variations with one-tap copy; following the phases produces the working app, tests, and a tagged v1.0.0 GitHub release. ## Quick Start Use the ai-message-polisher-builder skill to scaffold the message tone polisher web app with the 12-tone catalog, dual-variation output, and MVT branding.

Frequently Asked Questions about ai-message-polisher-builder

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I build an AI message rewriting web app?▼

Follow the phased workflow: initialize a React 19 + TypeScript + Tailwind + Express + tRPC project, add a server-side rewrite procedure with tone and style prompts, generate two variations in parallel with Promise.all, then build the tone grid UI, VariationCard outputs, tests, and GitHub setup.

How do I stop an LLM from answering a message instead of rewriting it?▼

Prepend a CORE_GUARDRAIL block as the first element of every system prompt. It instructs the model to treat input strictly as raw text to transform, never answering questions, resolving complaints, or adding commentary, and a final reminder enforces output of only the rewritten text.

How do I generate two LLM response variations at once?▼

Call the LLM twice in parallel using Promise.all with the same user message but different variation hints in the system prompt. Variation A uses slightly more direct phrasing and Variation B slightly warmer phrasing, returning both as a two-element array.

Can I add new tones to the message polisher?▼

Yes. Add the prompt string to the tone prompts map, add the enum value to the Zod schema, add a button entry to the tones array in the UI, and update the Help/FAQ tone list. The catalog currently ships with 12 tones and 3 style modifiers.

How should LLM calls be tested without real API requests?▼

Mock the invokeLLM function in Vitest tests so no real API calls occur. Cover valid tone and style combinations returning two string variations, whitespace-only input throwing a validation error, all 12 tones passing the Zod schema, and unknown tones being rejected.

Related Skills

Explore More Software Engineering Skills→
wigxelwigxel
3

convex-quickstart

Provision and configure Convex backend environments for web applications.

Official
Intermediate
atemagueratemaguer

expo-deployment

Deploy Expo applications to iOS, Android, and web release targets.

Community
Intermediate
indranilbanerjeeindranilbanerjee
726

prompt-test

Run formal A/B tests of prompt variations and select the winning variant.

Community
Advanced
json-choijson-choi

expo-dev-client

Build custom Expo development client apps for iOS and Android via EAS.

Community
Intermediate