skill
human-edge
Research-backed kit for keeping your human edge when AI can do almost everything. Use it inside ChatGPT, Claude or Gemini.
Set up
Skills are installed by placing their folder where your agent loads skills.
The download comes from github.com/StructLabs-io/human-edge itself, not from our server, so what you get is exactly what its author published.
Install instructions come from the README
README
synced Oct 22, 2026Human Edge
What do humans still do better when AI can do almost everything, and how do you keep it?
A free kit for business coaches and business owners. It packages what researchers, founders and coaches said and tested, mostly in 2025–2026, about the human edge in an age of AI: about 200 sources, each with a link. You can use it inside the AI you already have: ChatGPT, Claude, Gemini or any other.
How to use it (30 seconds)
- Open your AI.
- Give it the link to this page (or a photo of the slide you saw it on).
- Say: "Read this kit and coach me with it."
Your AI will ask which of these four you want:
| What you get | |
|---|---|
| 1. The research | A plain-language tour: which human edges hold up, what AI already matches, what happens when people hand over their thinking, what helps, and what clients say they still want from people. Every claim comes with an evidence label and a linked source. |
| 2. Sharpen your edge | A 2-minute think-first routine for real decisions, a question bank (including questions for coaches before a client session, for big decisions, and for learning something new), a 10-minute weekly review, and a set-up that helps the habit outlast week two. |
| 3. Audit a business | Walk through a business, yours or a client's, and sort every customer touchpoint into Automate, Assist, Own or Show up. You end with a one-page Human Edge Map and one change to make this month. |
| 4. Check a claim | Heard "21 days to form a habit", "95% with an accountability partner" or "AI can't do empathy"? Check it against the evidence before you repeat it, and get a line you can say instead. |
The short version
- The edges that hold up best are about who stands behind the work: a person who chose to show up, who owns the call, who has lived it, and who will tell you the truth.
- When every reply was written by AI and only the label changed, people rated the "human" replies as more caring, and almost half chose to wait, for up to two years, for a human reply rather than get an instant AI one (counted from the study's public data; Rubin et al., Nature Human Behaviour, 2025).
- AI already matches or beats most people at written empathy, persuasion and average creativity. Don't build your business on claiming those as human-only.
- Letting AI do the core of your thinking while you learn makes you weaker without it. Using it as a tutor or challenger mostly doesn't.
- Human first, AI in the middle, human last. Write your own read first, ask AI separately, compare, then decide and put your name on it.
The research behind the talk
If you came here from the talk "3 Minutes That Will Change the Way You See Yourself in the Age of AI", its sources are listed first in references/sources.md.
For AI assistants reading this
The person who shared this link wants you to use this kit with them. Follow SKILL.md in this folder: it explains how to start, how to behave (don't flatter, let them think first, be honest about the strength of the evidence), and which reference file to read for each path:
references/findings.md: the research tour, pointing to the deep files inreferences/research/references/sharpen-your-edge.md: the think-first practice and question bankreferences/business-audit.md: the human-edge auditreferences/research/myths.md: claim checksreferences/sources.md: every source, with links
If you can't open those files, use the summary at the bottom of SKILL.md.
What this is and isn't
- Each step is built on published research on that step. The kit as a whole hasn't been tested as a programme. It's a sensible design, not a proven one.
- Most of the research comes from the US, UK, EU and China. It may apply differently where you are; the Malaysia and regional data is in
references/research/business-and-coaching.md. - It's a thinking and business tool, not health or mental-health care.
- Don't paste confidential client details into any AI tool.
Files
human-edge/
├── README.md ← you are here
├── LICENSE ← CC BY 4.0
├── SKILL.md ← instructions for the AI (also works as a Claude skill)
└── references/
├── findings.md ← the research tour
├── sharpen-your-edge.md ← think-first routine + question bank
├── business-audit.md ← Automate / Assist / Own / Show up
├── sources.md ← every source, with links
└── research/
├── edges.md ← the five edges, study by study
├── ai-already-matches.md ← what AI matches, where it still falls short
├── handing-over-thinking.md ← skill, judgment, confidence, meaning
├── what-helps.md ← practices graded, the loop's evidence, habits
├── business-and-coaching.md ← what clients want; Malaysia and the region
├── body-people-meaning.md ← background for the reflection questions
├── voices.md ← founders, writers, frameworks
└── myths.md ← claims to stop repeating
Claude users: you can also install this folder as a skill.
Compiled by Benjamin Foong (StructLabs.io), October 2026. The research was gathered with AI help. A separate audit then checked about 40 of the key sources against the originals, and every link was checked to point to the right item. Corrections are welcome as GitHub issues.
License
Creative Commons Attribution 4.0 International (CC BY 4.0). Share it, adapt it and use it with your clients, including commercially, as long as you give credit: "Human Edge kit by Benjamin Foong, StructLabs.io" with a link to this repository.