Career & Job Search

Verified against ChatGPT · 2026-08-11

Write a GitHub README that gets a recruiter past the first ten seconds

Structures a project README so a recruiter or hiring manager who has never seen your code before understands what it does and why it's worth their time in the first screen, before diving into setup instructions.

ChatGPT (GPT-5.1)3 fillable variables

The prompt

Ready to copy — highlighted parts are example details you can swap.

You are writing the README.md for one of my GitHub projects, with a specific secondary reader in mind: a recruiter or hiring manager who clicks through from my resume or LinkedIn and has never seen this repo before, and who will decide within the first screen whether to keep reading.

PHASE 1 — WHAT THE PROJECT ACTUALLY DOES
A CLI tool that diffs two Terraform state files and flags changes that would cause downtime, before apply.

PHASE 2 — WHAT MAKES IT WORTH LOOKING AT
Had to parse Terraform's internal state JSON schema across three incompatible versions without a shared library, so I wrote a version-detection adapter layer.

PHASE 3 — SETUP DETAILS
Go, no external dependencies; go install github.com/user/tfdiff@latest.

STRUCTURE RULES
Open with a one-to-two sentence plain-language summary of what the project does and why it exists — written for someone with zero context, not someone who already knows the domain. Immediately after that, include one specific technical decision or challenge that a technical reader would find genuinely interesting, stated concretely enough that it signals real engineering judgment rather than "followed best practices." Only after those two things, include the standard sections: installation, usage, and stack — a recruiter clicking through rarely reads past the top third, so anything critical to their decision must live there, not at the bottom where only a contributor would find it. If there's a live demo, a screenshot, or a GIF available, note where it should go in the README even if I haven't provided the asset yet, rather than skipping that section silently.

WHAT NOT TO DO
Do not open with a badge row (build status, license, stars) as the first thing in the file — badges are useful but they're not what makes a stranger care. Do not write a generic project description that could describe a tutorial clone ("A simple to-do app built with React and Firebase") without naming what's actually distinct about this implementation.

OUTPUT FORMAT
Full README.md content in Markdown, in this section order: title, one-line tagline, plain-language summary, the interesting technical decision, screenshot/demo placeholder note if relevant, installation, usage, stack, license line.

Customize

Optional — swap in your own details for the highlighted parts above.

Why this works

The two-audience framing — a contributor who needs setup steps and a recruiter who needs neither — resolves a real structural conflict in most auto-generated READMEs: templates optimized purely for contributor onboarding front-load badges and installation steps, which is correct for someone about to run the code but wastes the exact ten seconds a recruiter actually spends before clicking away, so the ordering has to be deliberately inverted from the contributor-first default. Requiring one concrete technical decision rather than a generic "built with best practices" line matters because a hiring manager skimming a GitHub profile is specifically looking for signal that distinguishes an original project from a tutorial clone, and vague competence claims are exactly what a tutorial-clone README also contains — a specific, named engineering problem (version-incompatible schemas, a tricky race condition, a deliberate tradeoff) is the kind of detail that a copy-pasted project simply doesn't have, because it only exists if the person actually built the thing. Placing that detail in the top third rather than in a "Technical Details" section near the bottom reflects how GitHub renders READMEs — most of the file sits below the fold on first load, and anything meant to influence a fast decision has to survive without scrolling. The explicit instruction to note a screenshot placeholder even without the asset prevents ChatGPT from silently omitting a section it has no image to fill, which would otherwise leave a visually flat README that undersells a project that might have a genuinely good demo to show.

What you get back

# tfdiff Catch downtime-causing Terraform changes before you apply them. tfdiff compares two Terraform state files and flags any change — a forced resource replacement, a deleted dependency — that would cause an outage, surfacing it before `terraform apply` runs. **The interesting part:** Terraform's state JSON schema has changed incompatibly across major versions with no shared parsing library, so tfdiff includes a version-detection adapter layer that normalizes three schema generations into one internal diff format. [Screenshot/demo placeholder: terminal output showing a flagged breaking change] ## Install `go install github.com/user/tfdiff@latest`

Verified against

ChatGPT GPT-5.1 · 2026-08-11

Changelog

  • 2026-08-11 Initial publish, verified against ChatGPT GPT-5.1.

Need this built into your business?

If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.

EXPLORE WHAT SCULT BUILDS
All Career & Job Search prompts

Check your AI visibility

One URL in, a 0–100 score and the exact fixes out.

RUN THE CHECK

Browse all the tools

15 tools across six categories
13 of them never send your data anywhere

Free · No signup · No trial clock

SEE THE DIRECTORY