What These Skills Are

Each skill in this collection is a single [name]-skill.md file — the full instruction set that tells the AI how to behave, what to do, and how to do it

‍

The core of every SKILL.md is a section called Prompt / System Instruction — a block of text that tells the AI exactly how to behave for that skill. That's the piece you'll be working with most directly.

Using These Skills in Claude (claude.ai)

Claude is the native environment these skills were built for. There are two ways to use them.

Option A — Claude Projects (Recommended)

Claude Projects let you give Claude persistent instructions that apply to every conversation in that project. This is the best way to use these skills because you set it up once and it works every time.

‍

How to set it up:

‍

1. Go to claude.ai and create a new Project
2. Give the project a name (e.g., "Pre-Intake Prep" or "Employer Brand Tools")
3. In the project settings, find the Custom Instructions or Project Instructions field
4. Open the SKILL.md file for the skill you want to use
5. Find the section titled Prompt / System Instruction — it contains a block of text inside a code block (marked with ```)
6. Copy that entire block of text and paste it into the Custom Instructions field
7. Save the project

‍

Every conversation you start inside that project will now use that skill automatically.

‍

Tip: You can also upload supporting documents to a project — your Choosability Canvas context file, your company brief, your team brand context document. Any file you upload becomes available to Claude in every conversation in that project, so you don't have to paste your brand context every time.

Option B — Paste at the Start of a Conversation

If you don't want to set up a project, you can paste the skill instructions at the start of any conversation.

‍

How to do it:

‍

1. Open the SKILL.md file for the skill you want
2. Find the Prompt / System Instruction section
3. Copy the full block of text
4. Start a new conversation in Claude
5. Paste the instruction text as your first message, then immediately follow it with your actual request or content

‍

Example:

‍

[paste the full skill instruction block here]

‍

Here's the job posting I'd like you to analyze: [paste your content]

‍

This works fine for one-off use. The limitation is that you have to paste the instructions every time you start a new conversation.

Which skills work best as Projects vs. one-off conversations

Best as permanent Projects (you'll use these repeatedly):

‍

• Pre-Intake Meeting Prep
• Content Alignment Tester
• BS & Cliché Detector
• Persona Mirror
• Business Language Translator
• Choosability Canvas Formatter
• Company Brief Builder
• Team Brand Context Builder

‍

Fine as one-off conversations (you use these situationally):

‍

• Role Playbook
• Intake Transcript Processor
• Hiring Manager Profile Builder
• Talent P&L Builder
• Hiring Process Mapper
• Engagement Survey Extractor
• Value Translator

Using These Skills in ChatGPT

ChatGPT has a feature called Custom GPTs that works similarly to Claude Projects. It lets you create a version of ChatGPT with custom instructions pre-loaded.

Option A — Create a Custom GPT (Recommended)

How to set it up:

‍

1. Go to chat.openai.com and click Explore GPTs in the sidebar
2. Click Create in the top right
3. You'll see two tabs: Create (a conversational setup wizard) and Configure (manual setup). Click Configure.
4. Give your GPT a name and a short description
5. In the Instructions field, paste the full content of the Prompt / System Instruction block from the SKILL.md file
6. If the skill benefits from having your brand context always available, upload your context documents (canvas, company brief, etc.) in the Knowledge section
7. Click Save and set visibility to "Only me"

‍

You now have a custom GPT for that skill that you can return to any time.

‍

Important note: ChatGPT's Custom GPT feature is available on paid plans (ChatGPT Plus, Team, or Enterprise). If you're on the free plan, use Option B.

Option B — Paste as a System Message

If you don't have access to Custom GPTs, paste the skill instructions at the start of a conversation — the same approach as Claude Option B above.

‍

One difference with ChatGPT: If you're using the API or a tool that gives you access to a "System" message field (separate from the user message), paste the skill instructions there. This keeps the instructions separate from your actual request and tends to produce more consistent behavior.

Limitations in ChatGPT

• ChatGPT does not have persistent memory across conversations by default (unless you've enabled the Memory feature in settings). This means context from one conversation won't carry into the next.
• The skills that rely on web search (like the Pre-Intake skill fetching competitor job postings) require ChatGPT to have web browsing enabled. Check your settings to make sure this is on.
• File uploads work in ChatGPT, but the handling of long documents can be less reliable than in Claude. For context-heavy skills (like the Engagement Survey Extractor with thousands of responses), Claude tends to perform better.

Using These Skills in Gemini

Google's Gemini (available at gemini.google.com) can be used with these skills, though the setup is slightly different from Claude and ChatGPT.

Option A — Gemini Gems (Recommended)

Gemini has a feature called Gems — custom AI assistants with pre-loaded instructions — available on Google One AI Premium and Google Workspace plans.

‍

How to set it up:

‍

1. Go to gemini.google.com
2. In the left sidebar, find Gems and click New Gem (or Gem manager)
3. Give your Gem a name
4. In the Instructions field, paste the full content of the Prompt / System Instruction block from the SKILL.md file
5. If you want your brand context always available, you can include it directly in the instructions (paste the canvas or company brief text directly into the instructions field after the skill instructions)
6. Save the Gem

‍

Your Gem will now be available from the sidebar for any future conversation.

Option B — Paste at the Start of a Conversation

Same approach as Claude and ChatGPT Option B — paste the skill instructions at the start of a conversation, followed immediately by your actual request.

Limitations in Gemini

• Gem creation requires a paid Google plan. Free Gemini users will need to use Option B.
• Gemini's handling of very long inputs (like thousands of engagement survey responses) may be less consistent than Claude. For high-volume data skills, test with a smaller sample first.
• File upload support in Gemini varies by plan and interface. For skills that require document uploads (intake transcripts, canvas files, etc.), paste the content directly if uploads aren't working reliably.
• Some skills use web search to research competitor job postings and company data. Gemini has web access built in, but the depth and reliability of that research may vary compared to Claude's search capabilities.

General Tips for All Platforms

Read the instructions first. This document has all the instructions you need: what to bring to the conversation, what to expect, and how to get the most out of it. Two minutes reading the guide will save you frustration.

‍

Bring your context documents. Many of these skills work better when they have access to your Choosability Canvas, your company brief, or your team brand context. Either upload them to your project/custom GPT/gem, or paste them into the conversation alongside the skill instructions.

‍

Be specific in your inputs. These skills are built to work with real, specific material. The more concrete and detailed the information you bring, the more useful the output. Vague inputs produce vague outputs — every time.

‍

You can combine skills in a single conversation. For example: run the Value Translator on your intake output, then use those value statements to inform the Role Playbook in the same session. The skills are designed to work together.

‍

The skills don't remember previous conversations (unless you're using a Project, Custom GPT, or Gem with persistent context). If you're continuing work across multiple sessions, keep your key documents — canvas, company brief, intake forms — somewhere you can paste them in quickly.

‍

‍Estimates are always better than blanks. Several skills (especially the Talent P&L Builder) work with imperfect data. Don't wait until you have perfect information. Document your estimates, use the skill, and improve the inputs over time.

The Basic Idea

Every skill is a markdown text file. You could open it in a text editor and change it by hand — but you don't have to. The easier approach is to hand it to an AI, describe what you want to change, and let it make the edits and return a new version of the file.

‍

This works in Claude, ChatGPT, or Gemini. The process is the same in all three.

How to Do It

Step 1 — Start a fresh conversation

Open a new conversation in whichever AI tool you're using. Do not use a conversation that already has a skill loaded — you want a clean context so the AI is editing the skill, not running it.

Step 2 — Paste the skill file and explain what you want

Start your message with a brief framing sentence, then paste the full contents of the SKILL.md file, then describe your change. Like this:

‍

"This is a skill file I use with AI tools. It's designed to [one sentence on what it does]. I'd like to make some changes to it. Please read it carefully, then make the following changes:

‍

[describe your change or changes]

‍

Once you've made the changes, output the complete updated skill file so I can copy it and replace the original."

‍

Then paste the full SKILL.md content directly into the message.

Step 3 — Review the changes

The AI will return an updated version of the skill. Read through it — especially the section you asked to change — to make sure it reflects what you wanted. If something is off, ask it to adjust.

Step 4 — Save the updated file

Once you're happy with the changes, copy the full updated text and either:

‍

• Paste it back into your Project, Custom GPT, or Gem instructions to replace the old version
• Save it as a new SKILL.md file on your computer, replacing the original

Example Prompts for Common Changes

Making the tone warmer or more direct: "The tone of this skill feels a bit formal. I'd like it to feel more like a knowledgeable colleague having a conversation — less instructional, warmer. Please adjust the tone throughout, especially in the opening and in how it delivers feedback."

‍

Adding a new thing the skill should always do: "I'd like the skill to always ask for the hiring manager's name at the start of a session, so it can reference them by name throughout. Please add this to the opening behavior."

‍

Removing something the skill does that you don't want: "This skill currently offers to do a full rewrite at the end of its feedback. I'd like to remove that option entirely — it should only ever give coaching feedback, never offer a rewrite. Please remove any mention of the rewrite offer."

‍

Making it focus more narrowly: "This skill currently covers a lot of ground. I'd like to narrow it so it only focuses on [specific aspect] and ignores everything else. Please adjust the scope accordingly."

‍

Adding industry-specific context: "I work exclusively in the construction industry. Please adjust this skill so that all its examples, language, and framing are relevant to construction — not general corporate or tech contexts."

‍

Changing what the output looks like: "The skill currently produces a long, detailed report. I'd like it to produce a much shorter output — a maximum of five bullet points, plain language, no headers. Please restructure the output format accordingly."

A Few Tips

Be specific about what you want changed. "Make it better" is hard to act on. "Make the questions it asks feel more conversational and less like a form" is actionable.

‍

Ask for one or two changes at a time. You can make multiple changes in a single message, but if you ask for too many things at once the AI may lose track of some of them. Three or fewer changes per session is a good rule of thumb.

‍

Always ask for the complete file back. Don't let the AI show you just the changed sections — ask for the full updated file. Partial outputs make it easy to lose track of what changed and introduce inconsistencies.

‍

Keep a copy of the original. Before you replace a skill with an edited version, save the original somewhere. If the changes don't work the way you expected in practice, you'll want to be able to go back.

‍

Test before committing. After updating your Project, Custom GPT, or Gem with the new version, run a quick test conversation to make sure the skill behaves the way you intended. It takes two minutes and saves a lot of frustration.

When to Edit vs. When to Start Fresh

Edit the skill when:

‍

• The skill works well overall but one specific behavior needs adjusting
• You want to add something the skill doesn't currently do
• The tone or format isn't quite right for your context
• You want to narrow or expand the scope slightly

‍

Start with a fresh skill build when:

‍

• The skill's core purpose needs to change significantly
• You've made so many edits that the original structure no longer makes sense
• You're adapting a skill for a completely different use case or audience

‍

For a fresh build, describe what you want to the AI from scratch rather than editing the existing file. The result will be cleaner than a heavily-patched existing skill.

‍