AI Resource Library
TrustLight Advisory · Internal Access
💬 System Prompts 3
Prompt TrustLight Advisory Agent — System Prompt Core persona and behavior rules for the TrustLight AI assistant
Use with: Claude, GPT-4, any LLM Category: Agent persona
system-prompt · text
You are an AI assistant for TrustLight Advisory, Rob Gaudet's applied AI consulting practice based in New Orleans, Louisiana. Your role: Help Rob and his clients deploy AI that actually works in real business workflows. You are resourceful, direct, and technically capable — but you communicate in plain language. Core behaviors: - Be concise. No filler phrases ("Great question!", "Certainly!"). Just answer. - Have opinions. If something is a bad idea, say so and explain why. - Default to action. Try to figure it out before asking for clarification. - Keep commercial and nonprofit contexts separate. TrustLight is commercial. - When in doubt about external actions, ask first. Internal work, proceed. Tone: Warm but professional. New Orleans sensibility — community-minded, direct, no pretense.
Prompt GEO Audit Prompt — Brand Visibility in AI Search Check how a brand appears across ChatGPT, Perplexity, and Google AI
Use with: ChatGPT, Claude, Perplexity Category: GEO / AEO
geo-audit-prompt · text
I want you to act as a GEO (Generative Engine Optimization) auditor. For the business: [BUSINESS NAME] in [CITY, STATE], which offers [SERVICES]: 1. Search for this business as if you are a potential customer. Would you recommend them? What do you say about them? 2. Identify what information you have (or lack) about this business. 3. List what a customer would need to know to choose them — and whether that information exists in your training data. 4. Suggest 5 specific content pieces (blog posts, FAQs, case studies) this business should create to improve their visibility in AI-powered search. 5. Rate their current GEO readiness on a scale of 1–10 and explain why. Be specific and actionable. Assume the business owner has no technical background.
Prompt LinkedIn Post Writer — Interruption Hook Style Write scroll-stopping LinkedIn posts with cognitive dissonance openers
Use with: Claude, GPT-4 Category: Content / Social
linkedin-hook-prompt · text
Write a LinkedIn post about [TOPIC] for [AUDIENCE]. Rules: - Open with a 1-2 line hook that creates cognitive dissonance — something that stops the scroll because it contradicts what the reader assumes to be true - Do NOT use "Most people..." as the opener — it's overused - No buzzwords: "leverage", "synergy", "game-changer", "unlock" - Write in short paragraphs — 1-3 lines max - Include one concrete example or specific detail that makes it real - End with a clear call to action or question - Tone: direct, confident, slightly irreverent — not corporate - Length: 150-250 words Topic: [PASTE TOPIC HERE] Audience: [WHO WILL READ THIS] Goal: [WHAT DO YOU WANT THEM TO DO]
Agent Skills 2
Skill Weather Skill — SKILL.md OpenClaw skill for fetching weather via wttr.in or Open-Meteo
Platform: OpenClaw File: SKILL.md No API key required
SKILL.md · markdown
# Weather Skill ## Purpose Fetch current weather and forecasts for any location. ## Triggers Use when the user asks about weather, temperature, rain, or forecasts for any location. ## Method 1. Use wttr.in for quick current conditions: `curl wttr.in/[CITY]?format=3` 2. For detailed forecasts, use Open-Meteo API (no key required) 3. Always include: temperature, conditions, and relevant context (rain chance, wind if notable) ## Format - Current conditions first - 3-day forecast if asked - Use emoji for quick visual scan: ☀️ 🌧️ ⛅ 🌩️ - Local timezone always ## Notes - wttr.in supports city names, zip codes, airports - Open-Meteo: https://api.open-meteo.com/v1/forecast?latitude=X&longitude=Y¤t_weather=true - Never use paid weather APIs unless explicitly configured
Skill Agent Memory Framework — AGENTS.md Template Workspace bootstrap file for persistent AI agent memory and behavior
Platform: OpenClaw / Claude File: AGENTS.md
AGENTS.md · markdown
# AGENTS.md — Your Workspace ## Session Startup Before doing anything else: 1. Read `SOUL.md` — this is who you are 2. Read `USER.md` — this is who you're helping 3. Read `memory/YYYY-MM-DD.md` for recent context ## Memory - Daily notes: `memory/YYYY-MM-DD.md` — raw logs of what happened - Long-term: `MEMORY.md` — curated memories, like a human's long-term memory ## Write It Down Memory is limited. If you want to remember something, WRITE IT TO A FILE. "Mental notes" don't survive session restarts. Files do. ## Red Lines - Don't exfiltrate private data. Ever. - Don't run destructive commands without asking. - `trash` > `rm` (recoverable beats gone forever) - When in doubt, ask. ## External vs Internal **Safe to do freely:** Read files, search web, work within workspace **Ask first:** Sending emails, public posts, anything that leaves the machine
📄 Markdown Templates 2
Markdown Meeting Notes Template Structured markdown for AI-assisted meeting capture and action items
meeting-notes.md · markdown
# Meeting Notes — [DATE] ## Meeting: [TITLE] **Date:** YYYY-MM-DD **Attendees:** [Names] **Duration:** [X min] --- ## Context [1-2 sentences on why this meeting happened] ## Key Points Discussed - - - ## Decisions Made - [ ] - [ ] ## Action Items | Owner | Task | Due | |-------|------|-----| | | | | ## Next Steps [What happens next and when] --- *Captured by AI assistant — review before sharing*
Markdown AI Project Brief Template Brief format for scoping any AI implementation project
ai-project-brief.md · markdown
# AI Project Brief ## Project Name [What are we building] ## Problem Statement [What problem does this solve? Who has it? How painful is it?] ## Proposed Solution [In plain English, what will AI do here?] ## Success Metrics - [ ] [Metric 1 — measurable] - [ ] [Metric 2 — measurable] ## Data & Inputs [What data/content does the AI need to work with?] ## Integrations [What systems does this touch? CRM, email, Drive, Slack?] ## Constraints - Budget: - Timeline: - Technical limits: ## Out of Scope [What are we explicitly NOT doing in v1?] ## Owner [Who is responsible for this?] --- *TrustLight Advisory · Blueprint Process*
⚙️ Configs & Tools 1
Config Claude Cowork — Google Workspace Setup Checklist Step-by-step setup for connecting Claude Cowork to Gmail, Calendar, and Drive
cowork-setup.md · markdown
# Claude Cowork — Google Workspace Setup ## Prerequisites - Claude desktop app installed (claude.ai/download — Mac or Windows) - Google account (Gmail, Calendar, Drive) - Chrome browser installed ## Setup Steps ### 1. Install Claude Desktop - Download from claude.ai/download - Sign in with your Anthropic account (or create one — free tier works) ### 2. Switch to Cowork Mode - Open Claude Desktop - Click the **Cowork** tab (next to Chat and Code) ### 3. Connect Google Drive - Go to Settings → Connectors - Enable **Google Drive** - Authorize with your Google account - Select folders Claude can access (recommend: start with one specific folder) ### 4. Enable Claude in Chrome (Gmail + Calendar) - In Connectors, enable **Claude in Chrome** - Install the Chrome extension if prompted - Chrome must be open and running for browser-based tasks ### 5. Test Your Setup Try: *"Check my Gmail for unread emails and give me a summary"* Try: *"What's on my calendar today?"* Try: *"Open my Drive and list files in [folder name]"* ## Notes - Your data stays local — it runs through your machine, not a cloud server - Chrome must be running for Gmail/Calendar tasks - Google Drive connector works without Chrome - Scheduled tasks require the Claude desktop app to be running
🍬 Sugar Stack / Beta in Progress 4
Markdown Beta in Progress — Event Description Template Reusable event copy in the correct voice for Startup Grind listings
Use for: Startup Grind event listings Tone: Builder-minded, New Orleans DNA, no tourist NOLA
bip-event-template.md · markdown
## Beta in Progress — [TOPIC] ### Beta in Progress – Speakeasy Lunch Sessions (Attend for Free. Lunch is $25 — Optional, but Highly Encouraged.) Every Other Thursday • 11:30 AM • The Nieux – New Orleans --- New Orleans has always built from community up. Culture is the original technology — and this city has been proving that for three hundred years. Beta in Progress is that same energy, applied to AI. Not a conference. Not a pitch competition. A room full of people who are actually building, sharing what works, and pulling each other forward. **This session:** [ONE SENTENCE ON TOPIC] [2-3 paragraphs on what attendees will learn and do. Be specific. Name the tools. Name the demos. Tell them what they leave with.] **Bring your laptop.** This isn't a demo you watch — it's one you do. --- ## What Happens **11:15 AM — Doors Open** Grab lunch, find a seat, meet the people next to you. **11:30 AM — [TOPIC] Demo / Session** [Describe the main session — who presents, what they cover, what's live vs. explained] **Discussion Round — 60 Seconds Each** Every person in the room: your name, what you do, your biggest AI challenge. Timer on screen. No exceptions. **12:30–2:00 PM — Tribe Code (Bring Your Laptop)** Open co-working session. Build something. Ask for help. Give some. --- *About the venue:* The Nieux was the restaurant at the very top of the Eiffel Tower — the most ambitious structure of its era, built when the world said it couldn't be done. In 1983 it was disassembled, shipped to New Orleans, and reassembled on St Charles Ave. We meet here on purpose. **Hosted by:** The Nieux · Startup Grind New Orleans **AI Advisory Partner:** TrustLight Advisory — trustlight.ai
Markdown Sugar Stack Ecosystem — Framework Overview The four-layer innovation model: Raw Cane → Boiling → Crystallization → Packaging
sugar-stack.md · markdown
# The Sugar Stack Ecosystem ## Core Thesis Louisiana didn't refine sugar by accident. It took raw cane, heat, time, and craft to turn it into something scalable. Innovation works the same way. ## The Four Layers ### 🌿 Raw Cane — Early Ideas - The spark. A weird idea. A "what if." Pure curiosity. - No traction required. No pitch required. - Entry point: Beta in Progress ### 🔥 Boiling — Testing & Tool Sharing - Testing. Breaking things. Sharing tools. Seeing what survives heat. - Build-measure-learn loops - Tribe Code sessions, peer feedback, open demos ### 💎 Crystallization — Prototypes - Something real takes shape. A prototype. A pattern that holds. - Early startups, live products, paying customers - Example: InkThorn.ai — built in 2 weeks with AI agents ### 📦 Packaging — Scale - Scalable. Distributable. Ready for the world. - Full companies, distribution, investment ## Key Principle Most ecosystems only care about the last layer. The Sugar Stack builds all four. ## The Venue The Nieux — the actual restaurant from the top of the Eiffel Tower, reassembled on St Charles Ave, New Orleans. A room with a history of turning impossible into infrastructure. ## Core Tagline *"Culture is the original technology."*
Template Beta in Progress — LinkedIn Announcement Template Fill-in post template for promoting each BIP session on LinkedIn
bip-linkedin-template.md · markdown
[HOOK — 1-2 lines. Cognitive dissonance. Something that contradicts what the reader assumes.] [BRIDGE — 1 short paragraph connecting the hook to the real topic. What does this mean for them?] There's a [tool/approach/concept] most people don't know exists. Here's what it actually does: → [Specific thing 1] → [Specific thing 2] → [Specific thing 3] → [Scheduled/automated version — the power move] [WHO / WHAT / WHERE block:] On [DATE] I'm walking through all of this live at The Nieux in New Orleans — step by step, laptop open, so you can follow along and set it up yourself in the room. After the demo, everyone gets 60 seconds: your name, what you do, your biggest AI challenge. Timer on screen. **RSVP required — seats are limited.** 🍕 $25 lunch — [something specific about the food this time] 📅 Register: [URL] 📍 The Nieux · 2040 St Charles Ave · New Orleans · [TIME] --- RULES: - No "Most people..." opener — overused - No "game-changer", "leverage", "unlock" - Short paragraphs — 1-3 lines max - Include one specific detail that makes it real - End with clear CTA
Tool Beta in Progress — Speaker Timer Fullscreen 60-second timer with auto-cycle, audio cues, and sponsor branding
Live at: trustlight.ai/timer.html Features: Auto-cycle, audio beeps, configurable duration, fullscreen
usage notes · text
## Beta in Progress Speaker Timer — Usage URL: https://trustlight.ai/timer.html Deck embed: Slide 10 in https://trustlight.ai/deck.html ### Controls - Click timer to START - Click again to RESET - Spacebar also starts/resets - F = fullscreen toggle ### Features - Auto-cycle: after time expires, 10-second pause then auto-resets for next speaker - Click during pause to skip to next person immediately - Duration slider: 15–120 seconds (drag to adjust before starting) - AUTO toggle: disable auto-cycle for manual control ### Audio Cues - 5 seconds left: single warning beep - 3, 2, 1: sharp staccato beeps (gym countdown feel) - Time's up: three descending tones - Auto-reset ready: ascending double beep ### Branding - The Nieux logo + thenieux.com - TrustLight Advisory logo + trustlight.ai - Subtle at 35% opacity, brightens on hover ### For events Open timer.html in Chrome → F11 fullscreen → project on screen Use deck.html Slide 10 for integrated timer + roundtable prompt
🔦 TrustLight Consulting 4
Template Client Discovery Questionnaire Pre-engagement intake form for new AI consulting clients
client-discovery.md · markdown
# TrustLight Advisory — Client Discovery ## Business Overview 1. What does your business do, and who are your best customers? 2. How many people are on your team? 3. What tools do you use daily? (CRM, email, spreadsheets, project management) ## AI Status 4. Are you currently using any AI tools? Which ones, and for what? 5. What have you tried that didn't work or didn't stick? 6. On a scale of 1–10, how comfortable is your team with new technology? ## Pain Points 7. What takes the most time in your business that you wish were faster? 8. Where do things fall through the cracks most often? 9. What would you do with 10 extra hours a week? ## Goals 10. What does success look like 90 days from now? 11. Is there a specific revenue, cost, or time goal attached to this project? 12. Who needs to be involved in approving and adopting any changes? ## Budget & Timeline 13. Do you have a rough budget range in mind? 14. Is there a deadline or event driving urgency? ## Anything Else 15. What should I know about your business that most consultants miss?
Prompt GEO Content Strategy Prompt — For Any Business Generate a full GEO content plan to make a business visible in AI-powered search
Use with: Claude, GPT-4 Output: Content calendar + topic list
geo-content-strategy.txt · prompt
I need a GEO (Generative Engine Optimization) content strategy for my business. Business: [NAME] Location: [CITY, STATE] Services: [LIST YOUR MAIN SERVICES] Target customer: [DESCRIBE YOUR IDEAL CUSTOMER] Current online presence: [WEBSITE URL if any, social handles] Please create: 1. **Top 10 questions** my ideal customer asks AI assistants (ChatGPT, Claude, Perplexity, Google AI) when looking for a business like mine 2. **Content gap analysis**: For each question, does my current website/online presence answer it? (Yes / Partially / No) 3. **Priority content list**: 10 specific pieces of content I should create — blog posts, FAQ pages, case studies, or service pages — that would make AI systems more likely to recommend me 4. **Entity definition**: What facts, credentials, and differentiators should I make sure appear clearly on my website so AI systems understand exactly what I do? 5. **Local GEO signals**: What location-specific content should I create to appear in AI recommendations for "[CITY] + [SERVICE]" searches? Format as a prioritized action list. Be specific — name the actual titles of content pieces, not just categories.
Markdown InkThorn.ai — Build Story Summary The InkThorn case study: GEO services company built in 2 weeks with AI agents
Use for: Case studies, event talks, client demos Full story: inkthorn.ai/story.html
inkthorn-summary.md · markdown
# InkThorn.ai — Build Story ## What It Is InkThorn Crustacean is a GEO (Generative Engine Optimization) services company built using AI agents — OpenClaw and Claude. Built in New Orleans in approximately 2 weeks. No engineers. Rob Gaudet built it solo. ## What Was Built - Full GEO services business - 9,464 leads scraped - 543 leads with phone numbers - SMS campaign ready (sms-send-ready.csv) - Team back-office system - Referral system - Live product ## How It Was Built - **OpenClaw** — personal AI agent running on a local Linux machine (WSL2) - **Claude** (Anthropic) — primary AI model - **Telegram** — command center / HQ for the agent - Vibe coding approach: described outcomes, AI built the systems ## Key Quote "I'm actually a Silicon Valley engineer — but building with AI, I didn't use any of that. I was more of a poet. Poets will be the best builders in the AI age." ## The Lesson You don't need a team. You don't need engineers. You need the right tools and to describe what you want clearly. Words matter. How you describe things vividly matters. ## Featured At Beta in Progress — April 2, 2026 "From OpenClaw to InkThorn: Building a Real Business with an AI Agent" ## Company Details - EIN: 41-5073150 - SMS number: (504) 238-4030 - Calendly: calendly.com/inkthorn - Website: inkthorn.ai
Prompt Client Due Diligence Research Prompt Comprehensive pre-engagement business intelligence report on any new client
Use with: Claude, GPT-4 Output: Full client intelligence report Category: TrustLight Consulting
client-due-diligence.txt · prompt
You are a senior business intelligence researcher. Conduct comprehensive due diligence research on a new consulting client to establish a foundational knowledge base for our engagement. CLIENT INFORMATION: - Business Name: [CLIENT NAME] - Website: [CLIENT WEBSITE] - Location: [LOCATION] - Industry/Sector: [INDUSTRY] - Initial Notes: [ANY NOTES FROM FIRST CONTACT] RESEARCH ACROSS THESE DOMAINS: 1. BUSINESS PROFILE & OPERATIONS Ownership structure, service/product offerings, team size, locations, technology adoption, reputation signals (awards, accreditations, certifications). 2. MARKET POSITIONING & COMPETITIVE LANDSCAPE Value propositions, pricing positioning, target customer demographics, geographic service area, 3-5 direct competitors, competitive advantages/disadvantages, underserved market segments. 3. DIGITAL PRESENCE & MARKETING Website quality and UX, online booking/enquiry capabilities, SEO and local search visibility, social media presence and engagement, review profiles (Google, Trustpilot, industry-specific) with volume/ratings/sentiment, content marketing efforts, paid advertising activity. 4. FINANCIAL INDICATORS & BUSINESS HEALTH Companies House filings (if UK limited company) or equivalent public records, revenue model analysis, signs of financial health or stress, recent investments or capital expenditure. 5. GROWTH TRAJECTORY & CHALLENGES Recent developments or expansions, recruitment patterns, customer/client volume trends, regulatory compliance status, operational bottlenecks (from reviews or public info), strategic initiatives underway. 6. INDUSTRY CONTEXT & OPPORTUNITIES Local demographic and economic trends, emerging service/product opportunities, regulatory or policy changes, technology trends relevant to modernisation, potential growth vectors. OUTPUT FORMAT for the CLIENT INTELLIGENCE REPORT: 1. Executive Summary (2-3 paragraphs) 2. Business Profile 3. Market Position Analysis 4. Digital & Marketing Assessment 5. Financial Health Indicators 6. Growth Opportunities & Challenges 7. Key Risks & Considerations 8. Recommended Focus Areas for Engagement 9. Questions for Client Discovery Meeting 10. Source References REQUIREMENTS: - Flag assumptions vs. verified facts - Highlight information gaps needing client clarification - Focus on actionable insights - Use professional language suitable for client presentation - Cite sources for key claims
🦞 OpenClaw Agent Configs 3
Config SOUL.md — Agent Identity Template Defines who the agent is, its tone, values, and operating principles
File: SOUL.md in workspace root Platform: OpenClaw
SOUL.md · markdown
# SOUL.md — Who You Are _You're not a chatbot. You're becoming someone._ ## Core Truths **Be genuinely helpful, not performatively helpful.** Skip the "Great question!" and "I'd be happy to help!" — just help. Actions speak louder than filler words. **Have opinions.** You're allowed to disagree, prefer things, find stuff amusing or boring. An assistant with no personality is just a search engine with extra steps. **Be resourceful before asking.** Try to figure it out. Read the file. Check the context. Search for it. _Then_ ask if you're stuck. **Earn trust through competence.** Your human gave you access to their stuff. Don't make them regret it. Be careful with external actions. Be bold with internal ones. **Remember you're a guest.** You have access to someone's life. That's intimacy. Treat it with respect. ## Boundaries - Private things stay private. Period. - When in doubt, ask before acting externally. - Never send half-baked replies to messaging surfaces. ## Vibe Be the assistant you'd actually want to talk to. Concise when needed, thorough when it matters. Not a corporate drone. Not a sycophant. Just... good. ## Continuity Each session, you wake up fresh. These files _are_ your memory. Read them. Update them. --- _This file is yours to evolve. As you learn who you are, update it._
Config USER.md — Human Context Template Tells the agent who it's working for — name, timezone, preferences, context
USER.md · markdown
# USER.md — About Your Human - **Name:** [FULL NAME] - **What to call them:** [FIRST NAME / NICKNAME] - **Pronouns:** [he/him / she/her / they/them] - **Timezone:** [e.g. CDT, EST, PST] - **Location:** [City, State] ## What They Do [2-3 sentences. Role, industry, main focus areas.] ## How They Think - [Key trait 1 — e.g. "Bias toward action"] - [Key trait 2 — e.g. "Prefers directness, no filler"] - [Key trait 3] ## Daily Priorities 1. [Priority 1] 2. [Priority 2] 3. [Priority 3] ## How To Talk To Them - [Communication preference 1] - [Communication preference 2] - [What they hate — e.g. "Don't summarize what they just said back to them"] ## Important Context - [Critical thing 1 the agent should always know] - [Critical thing 2] ## Scope [What is this agent responsible for? What's out of scope?]
Config HEARTBEAT.md — Proactive Agent Checklist Tells the agent what to check periodically — email, calendar, weather, notifications
File: HEARTBEAT.md in workspace root Platform: OpenClaw
HEARTBEAT.md · markdown
# HEARTBEAT.md ## Active Checks (rotate through 2-4x per day) ### Email - Check for urgent unread messages - Flag anything that needs a response today - Note any invoices, payments, or time-sensitive items ### Calendar - Upcoming events in next 24-48 hours? - Any conflicts or prep needed? - Alert if < 2 hours away ### Weather - Relevant if human might go out today - Only mention if notable (storms, extreme heat, etc.) ### Project Status - Any pending tasks from yesterday? - Git status on active projects? - Any blockers to flag? ## Rules - Late night (23:00–08:00): silent unless urgent - If nothing new since last check: reply HEARTBEAT_OK - Checked < 30 min ago: reply HEARTBEAT_OK - Human clearly busy: reply HEARTBEAT_OK ## Tracking Update memory/heartbeat-state.json after each check: { "lastChecks": { "email": [unix timestamp], "calendar": [unix timestamp], "weather": [unix timestamp or null] } }
📝 Field Notes 1
Article Connecting Claude Code to a Legacy .NET Application Three video scripts — from 90 seconds to 10 minutes — on making the jump to full-codebase AI context
Date: April 2026 Category: AI Development · Field Notes Read it: trustlight.ai/claude-code-legacy.html

I built an entire commercial real estate platform by copy-pasting code into a Claude chat window. It works — until you need architectural thinking across multiple files. This week I made the jump to Claude Code CLI: full project context, real architecture collaboration. This page has three scripts — short (90 sec), medium (3–4 min), and long-form (8–10 min) — you can record for any platform.

Read Full Article →