AI Skill of the Day Digest — May 2026 (Part 2)

Every day, The Neuron teaches readers one practical AI skill they can use immediately. This second May collection picks up where Part 1 left off, with every published AI Skill from May 14 through May 31.

Skim the headings, grab the prompts that fit your work, and come back when you need a better way to use Claude, ChatGPT, Gemini, Codex, or an AI agent.

And yes, this is from May of 2026, but we published it in August 2026. You can think of this as a "straight from the archives" release!

How to use this digest

• Skimming? Each entry starts with the practical outcome.
• Trying one? Copy the included prompt or workflow and replace the bracketed details.
• Catching up? The skills are ordered by their original newsletter date.

🎓 May 14

Stop Saving Prompts in Docs. Use Claude Skills Instead

If you've been copy-pasting your best prompts from a Google Doc every time you need them, YouTube creator Alex Pereira just showed a better way.

Claude has a feature called Skills — reusable instruction sets you install once and trigger with a single slash command, in any chat. Think of it like saving a prompt permanently to Claude's brain. Alex built six of them to run his entire content workflow: hook generator, image creator, video generator, thumbnail builder, face lock (for consistent face generation), and caption writer — all fired with one word.

How to set one up:

• In Claude, click Customize → the + button → Create Skill

• Structure it in three parts: (1) a description + when Claude should trigger it, (2) your instructions, (3) examples of good output

• Save it — then type /[skill name] in any chat to activate

Alex's rule for when to bother: the task has to be recurring, structured, and have a consistent output format. Passes all three? Build it. Passes one or two? Keep it as a saved prompt.

`When I type /[skill name], do the following: Instructions: [What you want Claude to do]
Format: [Output shape — e.g. 10 hooks, numbered list, under 15 words each]
Voice rules: [Any style/tone requirements]
Examples of good output: [Paste 2-3 examples]
`

🎓 May 15

Design Goals and Hooks for Normal Work

The best agent workflows have two ingredients: a goal that defines "done" and hooks that keep the agent consistent while it works.

A goal is the finish line. A hook is a just-in-time rule that runs at the right moment. Developers use these for code checks, tests, and project rules. You can use the same pattern for normal work: trip planning, research, inbox cleanup, budget reviews, meeting prep, and anything else that requires multiple steps.

Copy this:

Help me turn this task into a long-running agent workflow: [TASK] Define: 1. The goal: what counts as done, in concrete terms. 2. The checkpoints: where the agent should stop and ask me. 3. The hooks: rules that should run automatically at the right moments. 4. The risky steps: anything public, expensive, irreversible, sensitive, or customer-facing. 5. The progress updates: what I should see while the work is running. Make it useful for a normal person, not a developer. Give examples of hooks I can use for this exact task. Favorite insight: agents get useful when they know when to continue, when to check themselves, and when to bother you.

🎓 May 17

Run an AI Security Gut Check

You do not need to be a security engineer to ask better security questions.

Today's skill: use AI to run a defensive "gut check" on any workflow where software, customer data, accounts, or automation are involved. The goal is simple: find the places where trust is being assumed.

This is useful because this week's biggest security stories were all about assumptions. A package install assumed the code was safe. A login system assumed a user with partial access could be trusted. A security team assumed old scanning methods could catch new attack patterns.

Copy this into ChatGPT or Claude:

`Act as a defensive security reviewer. Review this workflow, system, or automation:
[PASTE DESCRIPTION] Look for places where the system assumes trust too early. Focus on:

1. User accounts and permissions
2. Third-party packages or integrations
3. API keys, tokens, and credentials
4. Automated actions that could cause damage
5. Data that should stay private
6. Approval steps before anything public or irreversible
7. Monitoring logs I should check regularly For each risk, explain:

• What could go wrong
• Why a normal check might miss it
• The safest practical fix
• Whether this needs an expert review Keep this defensive only. Do not provide exploit steps.

`

🎓 May 18

Run a Platform Risk Audit

Here's the lesson from the Apple / OpenAI mess from last week: distribution is great until someone else owns the switch.

Today's skill is a platform risk audit. Use it anytime your product, workflow, business, or audience depends on another company's platform. That could mean Apple, Google, Amazon, Meta, TikTok, Slack, Salesforce, or even ChatGPT itself.

The goal is simple: find where you're exposed before the platform changes the rules, buries your feature, launches a competitor, or signs the same deal with someone else.

Copy this:

`Act as a platform strategy advisor. Audit this dependency:
[DESCRIBE PRODUCT, WORKFLOW, PARTNERSHIP, OR DISTRIBUTION CHANNEL] Find my biggest platform risks across: 1. Access: What could the platform block, limit, or delay?

1. Visibility: Could the platform bury us in the interface?
2. Economics: Could pricing, fees, or revenue share change?
3. Competition: Could the platform launch or favor a rival?
4. Data: What data do we need that the platform controls?
5. Switching costs: How hard would it be to move elsewhere?
6. Leverage: What do we control that the platform still needs? Return: - Top 5 risks

• Early warning signs
• A practical mitigation plan
• The one dependency I should reduce first Favorite insight: partnerships feel safest right before the platform reminds you who owns the door.

`

🎓 May 19

Make a permission map before connecting AI to anything sensitive

OpenAI’s personal finance launch is a good reminder: before you connect an AI tool to sensitive data, make it write you a permission receipt.

A permission map is a basic checklist of what the AI can see, what it can change, what it stores, and how you disconnect it. The trick is forcing the AI to separate read access from action access. Reading your bank transactions is one risk. Moving money, changing subscriptions, or storing private details in chat history is another.

Use this before connecting email, files, calendars, CRMs, bank accounts, or workplace tools (turn it into a skill with the @skill-creator-skill that you can call at any time):

You are my AI permissions auditor. Before I connect this tool or account, create a permission map. Tool/account I’m considering connecting: [paste tool] Data involved: [banking / email / calendar / files / CRM / other] (if not provided, look up the connector for me and find out) Give me: 1. What the AI can read. 2. What the AI can change or trigger. 3. What data may be stored in chat history, logs, or third-party systems. 4. What could go wrong in a normal workday. 5. The safest way to use it. 6. The exact steps I should take to disconnect or delete access later. Write this for a non-technical person. Be specific and practical; don't scare me away from using it, show me how to be as safe as possible when using it.

🎓 May 20

Use Gemini Like a Task Router to Help Best Use Gemini

Google now has Gemini everywhere. So the skill today: make Gemini pick the right surface before you start. Google’s own prompt guide says Gemini works best with direct, structured instructions, so give it a routing job first.

`I want to use Gemini for this task: [task]. Route me to the best Google tool, then give me:

1. The safest first prompt
2. The files or inputs to attach
3. What should stay read-only
4. What success looks like
5. One follow-up prompt

`Good rule of thumb: Start with reversible tasks: monitor, draft, compare, storyboard, or prototype before you let an agent send, buy, delete, or publish.

🎓 May 21

Turn Claude Into Your Full-Time Content Team

If you make content for a living (or want to), Modern Millie's 23-min tutorial is worth every minute. Here are the four workflows that actually move the needle:

1. Free profile audit. Install the Claude Chrome extension, open your Instagram or YouTube Studio analytics, then ask: "Audit my profile. What's working, what isn't, and what's the single most important thing I should change right now?" Claude reads the live page and gives you a full breakdown — hook strength, content gaps, growth opportunities.

2. 30-day content calendar in one prompt. Inside a Claude Project loaded with your brand info, paste this:

Based on everything you know about my brand and goals, build me a 30-day content calendar. Alternate between my content pillars and include 3 hook options per video.3. Full video script from the calendar. Pick a day you like, then:

Let's work on Day [X]. Use Hook [#]. Turn this into a full script with text hook, verbal hook, visual hook, talking points, and CTA.4. One video → one week of content. Get your YouTube transcript from transcript.io, start a new Project called "[Brand] YouTube to Everywhere," and paste: "Here's my transcript. Turn this into a full week of content." Out comes short-form scripts, captions, carousel outlines, and an email newsletter — all from the video you already made.

🎓 May 22

Give Codex a Definition of Done

Long agent tasks fail when the AI forgets what “done” means. Codex’s /goal mode fixes that by giving it a persistent objective it can keep checking as it works.

Use this for tasks with many steps: migrations, refactors, audits, bug sweeps, or report generation. The trick is to write the goal like a mini contract: outcome, constraints, and tests. If /goal does not appear, OpenAI says you can enable features.goals in config.toml or run codex features enable goals.

Try this:

`/goal
Audit this project for newsletter draft readiness. Definition of done:

1. Every section has the required header.
2. Every hyperlink is attached to a short, natural anchor.
3. No Treats or Around the Horn bullets use bold text.
4. Every technical term has a plain-English parenthetical on first use.
5. Return a short report with pass/fail status and exact fixes made. Before editing, make a checklist. After editing, run the checklist again and show me what changed.`

🎓 May 24

Let AI Write the Code That Emails Your Reports Automatically

If you still manually save reports as PDFs and email them one by one, Leila Gharani has a better way and it doesn't require a developer, a VBA macro, or Power Automate.

The trick is using Office Scripts (Excel's built-in automation tool that works in both desktop Excel and Excel online) paired with AI to write the code for you. The result: one click sends personalized PDF reports to every manager on your list, with only their data, automatically.

Here's how to set it up:

• Build your index table. Create a sheet with three columns: manager name, email address, and which sheet(s) belong to them. This is the list your script will loop through.

• Ask AI to write the script. Go to the Automate tab in Excel, open a New Script, then paste this into ChatGPT or Claude:

`Write an Office Script for Excel that does the following:

• Reads a table called "SendList" with columns: Name, Email, Sheets
• For each row, hides all sheets except the ones listed in the Sheets column
• Converts the workbook to a PDF
• Sends the PDF to the email address in that row with the subject "Your Report" and body "Hi [Name], please find attached your report."
• Unhides all sheets after sending
• Skips any row where the email or sheet name is missing or invalid Use TypeScript. Include error handling for missing sheets.`

• Paste the output into Office Scripts and run it. Add it as a button on your index sheet for one-click sending.

The AI does the hard part. You just describe what you want, paste the code, and click run. Leila's full walkthrough (including a downloadable file with the code) is worth a watch if you want to see it built step by step.

🎓 May 25

Run a “Sameness Detector” pass

Most AI drafts are fine. Fine is the danger.

As Dan Shipper argues, AI makes yesterday’s competence cheap, which means everyone can suddenly produce decent work. The problem is that decent work starts to repeat itself: same ideas, same examples, same section flow, same rhythm, same “smart” conclusion.

So today’s skill is running a Sameness Detector pass before you publish. You’re asking AI to find low-perplexity work, meaning the parts that feel too predictable. This includes repeated ideas, duplicate sections, overused phrasing, flat transitions, obvious examples, and “polished” paragraphs that add no new information.

The goal is not to make the draft weird. It’s to make it specific. Ask the AI to diversify the draft across ideas, structure, phrasing, evidence, examples, sentence rhythm, and emotional beats.

`Run a Sameness Detector pass on this draft. Your job is to find every place where the work feels predictable, repetitive, generic, or too similar to common AI output. Audit for sameness across:

1. Ideas: repeated points, obvious claims, or concepts that do the same job
2. Structure: sections that follow the same setup, order, or argument shape
3. Phrasing: repeated words, sentence patterns, transitions, and summary lines
4. Examples: generic examples that could apply to anyone
5. Evidence: unsupported claims or proof points that feel interchangeable
6. Rhythm: paragraphs or bullets that sound too similar in length or cadence
7. Emotional beats: places where the tone stays flat, over-polished, or predictable
8. Usefulness: sentences that sound smart but do not help the reader decide, act, or understand Then create a diversification plan:

• What should be cut?
• What should be combined?
• What needs a sharper example?
• What needs a more surprising angle?
• What needs more specificity from the audience, company, moment, or situation?
• Where should the structure change so the draft has more “perplexity,” meaning useful surprise and variation? Finally, rewrite the draft so it feels more specific, varied, and alive without making it longer by default. Audience:

[paste audience] Goal:
[paste goal] Draft:
[paste draft]`

🎓 May 26

Stop Asking Which AI Is Best. Ask Which One Fits Your Workflow.

Most AI debates miss the point. The question isn't "Copilot vs. Gemini vs. Claude." It's "which one lives where you already work?"

Patrick Giwa laid out a clean framework for this, and it's more useful than any benchmark:

• Use Copilot if your team runs on Microsoft 365. It's native inside Word, Excel, Outlook, Teams, and GitHub, so it can generate reports, summarize meetings, automate spreadsheets, and draft proposals without you ever leaving the app you're already in. Bonus: many enterprise companies block ChatGPT but allow Copilot, making it the most-adopted AI tool in corporate settings whether anyone admits it or not.

• Use Gemini if your work lives in Google Workspace. Gmail, Docs, Sheets, Drive: Gemini is built into all of it. Best for summarizing email threads, drafting slides and reports, and handling the async collaboration and meeting prep that knowledge workers spend half their day on.

• Use Claude when the task requires real thinking across large amounts of material: legal review, research synthesis, long-document analysis, or anything where you need the model to reason carefully rather than just execute quickly. It's not the default enterprise assistant, but it's the specialist you want for heavy lifting.

Patrick's actual point, and it's a good one: "the best AI isn't always the most popular one." It's the one that integrates into how your team already works. Routing the right task to the right model is itself a skill, and most people aren't doing it.

🎓 May 27

Map Who Gets the Upside

Before you pitch any AI change at work, map who gets the upside. This is the fastest way to find the person who will quietly block your “efficiency” project because, to them, it looks like more work, more surveillance, or fewer jobs.

Use a stakeholder upside map: a simple scan of everyone affected by a change, what they gain, what they lose, and what would make the deal feel fair. It is especially useful for automation, new agents, data projects, and team-wide AI rollouts; things your colleagues, staff, or citizens might naturally be opposed to if they don’t see the benefit.

Copy this prompt and run it before you launch a project to help you gain perspective and better bargaining chips so you can gain buy-in from other stakeholders: 

`I’m considering this AI project: [describe the project].
Create a stakeholder upside map with these columns:

1. Stakeholder group
2. What they gain if this works
3. What they might fear or lose
4. What proof would make them trust it
5. What ownership, participation, or benefit-sharing mechanism would make the change feel fair
6. What are the downsides if this goes wrong (and how we can mitgate them through preparedness) 7. What's the first conversation I should have with each stakeholder.

Be concrete. Include employees, managers, customers, legal/compliance, IT/security, and any affected community or partner group. Consider both their legitimate upside and downside in order to help me see the full impact of this project and navigate it's challenges as well as communicate it's benefits.`Resistance often shows up as “people hate change,” but a lot of the time, people hate being handed someone else’s upside as their downside. 

Now, uno-reverse that! Before you pitch a new AI workflow, ask your chatbot to argue against it from the perspective of the person most likely to be affected. You will usually find the real blocker faster: extra review work, unclear accountability, job insecurity, data exposure, or the fear that “efficiency” means fewer humans later.

`Act as the person most likely to oppose this AI project: [describe it].
Write the strongest good-faith argument against it.
Focus on risks, incentives, hidden costs, trust issues, and what (if anything) would make you support it.`Oh, and if you want to take this a step further to consider every ethical aspect of an initiative before you launch it, you could copy the above prompt along with the Pope’s Magnifica Humanitas doc and ask the AI to “help me create a prompt like this one I’m sharing, using the Magnifica Humanitas doc, to ask questions that aim to address each of the ethical points the doc presents to build an “ethical responsibility map" before launching a new initiative.”

You might not be able to answer every ethical question on every project you pursue, but hey, at least you’ll have taken the step to think about it… more than can be said about a lot of projects TBH.

🎓 May 28

Write an Agent Permission Brief

Before you hand an AI agent a login, credit card, CRM, or production system, write a permission brief.

A permission brief is the agent version of giving a new hire a role, a budget, and a manager. It defines what the agent can do, what it must ask before doing, where it should leave logs, and what happens when it fails.

This is especially useful after the Robinhood news, because the risk is no longer theoretical. Agents can now touch money. Soon, they will touch procurement, customer accounts, legal docs, and internal systems.

Use this prompt before connecting any agent to real data or real spending:

`You are my agent-risk reviewer. I am considering giving an AI agent access to [system/tool/account]. Create a one-page permission brief with:

1. Actions the agent is allowed to take.
2. Actions the agent must ask approval for.
3. Actions the agent is never allowed to take.
4. Spending, data, or customer-impact limits.
5. Required logs and where they should live.
6. Failure scenarios and rollback steps.
7. A first-week test plan using low-risk tasks. Ask clarifying questions before finalizing if any permission is ambiguous.`

🎓 May 29

Use Claude Code’s workflow mode for big messy tasks

Some AI tasks fail because they’re hard. Others fail because they’re too big for one chat window to hold in its tiny little model brain.

That’s where Claude Code’s new dynamic workflows come in. A workflow lets Claude write an orchestration script (a repeatable plan in code), then spin up subagents (smaller Claude workers) to tackle pieces of the job in parallel.

Use this for work where “check one thing” has secretly become “check 400 things.” Anthropic suggests codebase audits, large migrations, and research that needs cross-checking. Cat Wu shared a great example: cataloging hundreds of A/B test flags and finding stale ones set to 0% or 100%, in parallel, instead of one by one.

How to use it:

• Start in Claude Code.

• Use the word “workflow” in your prompt.

• Keep the first run tightly scoped because this can burn tokens fast.

• Ask Claude to verify findings before reporting them.

• Save successful workflows with /workflows so your team can rerun them.

Try this:

`Create a workflow to audit [specific folder, repo, docs set, or dataset] for [specific issue]. Before running it, show me:

1. The stages of the workflow
2. What each subagent will inspect
3. How findings will be verified
4. Any files or commands you plan to touch
5. The smallest safe first pass Start with a scoped sample first. Do not make changes until I approve the full workflow plan.

`

🎓 May 31

Connect ChatGPT to vidIQ and Get Better Insights for Your Channel and Content

ChatGPT gives generic YouTube advice because it only knows what you tell it. Connect it to your actual channel data and the advice gets a lot more useful, fast.

In this 7-minute video, the vidIQ team shows how to connect vidIQ (a YouTube analytics tool) to ChatGPT so it can pull your real channel stats, not just give you guesses based on general knowledge. The analogy ChatGPT gave when asked about the difference: "Without real data, I'm coaching a football team using inspirational quotes. With it, I have the playbook, the game film, and the live scoreboard open at once."

Here's how to set it up:

• Open ChatGPT, go to Settings > Apps > Browse Apps

• Search for vidIQ, hit connect, and link your YouTube account

• Now ChatGPT can see your actual channel data when you prompt it

Then run these three prompts in order:

`Prompt 1 — Find what's working in your niche:
For my niche, identify 5 topics with strong search demand and low to medium competition right now. For each, give me the keyword, search volume tier, competition score, and a specific video angle a smaller channel could win with. Format as a table. Prompt 2 — Analyze your outlier videos:
Pull my channel's top outlier videos from the last 6 months. Analyze what they had in common: title structure, topic, format, length. Give me 3 strategic patterns I can replicate in my next videos to increase the chance of another outlier. Prompt 3 — Get your next 5 video ideas:
Combine the insights from the previous two requests with what's currently working on YouTube right now to give me compelling titles and approaches for my next 5 videos.` The key insight: most people use AI like a vending machine ("give me titles"). This workflow turns it into a strategy system built on your actual data.