Video summary

The 7 ChatGPT Work Skills I Use Every Day (Copy Them)

Main summary

Key takeaways

Technology

Summary of the Video (Tech Concepts + Product Features)

The video explains how to use “Skills” in ChatGPT desktop and ChatGPT Workflows to generate more consistent, reusable, higher-quality outputs—without re-explaining the same setup every time. Skills are described as repeatable instructions that “lock in” your preferred workflow.

Core Setup: How Skills Work

Where to find Skills

  • In the ChatGPT desktop app, go to Plugins → Skills to view installed skills.
  • Some skills may be preloaded, while others are personal.

How to create/add a new skill

The recommended method:

  1. Start a new chat in ChatGPT Work
  2. Describe the skill/task you want
  3. ChatGPT saves it as a reusable skill you can call later

How to invoke skills

  • Skills are called using slash commands (e.g., /visualize, /lmk, etc.).

The “7 Work Skills” (Plus a Bonus) Used Daily

1) /visualize (Turn responses into interactive visuals)

  • Converts ChatGPT responses into visual dashboards/charts rather than plain text.
  • Example: visualizing personal budgeting data with an interactive monthly/annual view, including tables for costs and expenses.
  • Key benefit: easier interpretation and the ability to explore data interactively.

2) /lmk (“let me know”) (Push notifications when a task finishes)

This skill addresses the issue where tasks can finish, but you don’t receive notifications on mobile.

Workflow:

  • Create a skill that triggers a push notification via a third-party service.
  • The example uses Pushover.

Setup details (as described):

  • Install and log into Pushover to get a user key
  • Create an application in Pushover to get an API token
  • Provide the token/key to ChatGPT so it can send notifications

Invocation: /lmk

Automation example:

  • A Mac mini automation researches trending topics (e.g., from X/Twitter) and notifies the phone when attention is needed.

3) Handoff Skill (Port progress to another AI)

Use this when you want to continue work in another model/tool (e.g., Claude) without re-explaining everything.

The handoff includes:

  • the project goal
  • what’s already been done
  • decisions made
  • relevant files/links
  • the next steps

Example described:

  • An iPhone app project hit limitations in ChatGPT.
  • The creator generated a handoff, pasted it into another AI, fixed issues, then brought the updated handoff back.

4) Ask Skill (Delegate browser tasks to another AI)

Use this when you don’t need a full handoff—just fetch help from another AI.

How it works:

  • Integrates browser-based access, letting ChatGPT:
    • open a site for another model (e.g., grock.com / Grok)
    • ask questions there
    • return results into the main workflow

Examples:

  • Researching trending AI news topics when ChatGPT lacks real-time X access:
    • use Grok to access X-related information
  • Generating images with Gemini instead of ChatGPT for different image outputs

Invocation pattern shown:

  • A command like /ask (the subtitles indicate browser-forwarding/routing intent).

5) GrillMe Skill (Interview-first to prevent wrong assumptions)

This skill prevents ChatGPT from “taking off” and building before fully understanding the request.

It forces an interactive interview before execution:

  • uncovers assumptions
  • confirms decisions
  • gathers missing details
  • blocks action until you agree it’s ready

Notes:

  • The original idea is credited to “Matt PCO”, but the creator provides a scaled-down version to move faster.

6) Multi-step Workflow Skills (Example: sponsor video review)

These skills chain together repeated steps so you don’t redo them manually.

Example: “sponsor video review” workflow

  • Takes a video link from the creator’s Obsidian workspace (where notes/outlines live)
  • Uses browser actions to go to Vimeo
  • Drafts an email to the brand including the video link
  • Updates a schedule/status entry in Obsidian
  • Prompts you to review the draft in Gmail and send

Key technical idea:

  • Multi-step skills can combine browser use, Gmail integration, and local file access via Obsidian.

7) Hatch / Pet-related Skills (Bonus included)

Bonus “pet” feature:

  • /pet spawns an on-screen pet in the desktop app that stays visible and shows task status.

Bonus /hatch skill:

  • Creates custom pets from a text description and optional reference images.
  • Pets can be managed in Settings → Pets.

Bonus: Remotion Plugin Skill Package (Motion Graphics)

The final showcased “skill” is really a skill package included with a plugin:

  • Install the Remotion plugin
  • Remotion is said to come prepackaged with 12 skills (examples mentioned include best practices, captions, docs, maps, etc.)

How it helps:

  • You don’t need to know the internal skill names
  • When you request something like “create a motion graphic,” ChatGPT selects the relevant plugin skills automatically

The video creator notes they use Remotion skills in YouTube videos.

Reviews / Guides / Tutorials Highlighted

  • A step-by-step guide for creating/installing skills in the ChatGPT desktop app (Plugins → Skills)
  • A tutorial-style setup for Pushover (tokens/keys) to make /lmk work
  • Examples showing skills that:
    • visualize structured data
    • notify you automatically
    • hand off to other AIs
    • query other AIs via browser routing
    • prevent premature execution (GrillMe)
    • run multi-step tasks across tools (Vimeo/Gmail/Obsidian)

Main Speakers / Sources (End of Video)

  • Main speaker: the YouTube creator (unnamed in the subtitles; author of the skills/examples)
  • Source mentioned: Matt PCO (creator of the original GrillMe skill; the video creator made a scaled-down version)
  • Third-party tools/services referenced:
    • Pushover, Vimeo, Obsidian, Gmail
    • Claude / Gemini / Grok (via grock.com)
    • Remotion plugin

Original video