Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Write a clear, well-structured README for a software project or open-source repo. Use when asked to write or improve a README, document a project, or make a repo approachable. Produces a complete README — one-line pitch, badges, quickstart, usage, install, contributing, license — that gets someone from landing to running fast.
.claude/skills/mohitagw15856-readme-writer/SKILL.md| Model | Eval pass | Runs |
|---|---|---|
| gemini-3.6-flash | 100% | 2 |
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -22% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 32% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -32% | 0% |
| case-11 | ✓→✗ | ▼ Worse | 0% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | 15% | 0% |
The README is a project's front door — most people decide in seconds whether to use or bounce. This skill writes a clear, scannable README that answers what is this, why should I care, how do I run it immediately, then layers in the detail. Structured so a newcomer gets to a working result fast.
Ask for these only if they aren't already provided:
A complete README.md:
> One-line pitch — what it does and for whom.
(Badges line — build, version, license — as placeholders to fill.)
Why project]? — 2–3 sentences or bullets: the problem and what makes this worth using (honest, specific).
Features — the handful that matter, as a tight bullet list.
Quickstart
# install
# minimal working example…with the expected result shown.
Usage — the common cases, with short code examples. Link out to full docs rather than inlining everything.
Installation — fuller install/requirements if the quickstart was minimal.
Contributing — how to contribute / link to CONTRIBUTING; be welcoming.
License — the license line.
(Adapt sections to the project; omit what doesn't apply. Keep it scannable with clear headings.)
Open-source README best practices (one-line pitch, time-to-first-success quickstart, scannable structure, standard sections).
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 26,759 | 21,970 | -18% | 1 | 1 | 0% | 3,861 | 3,049 | -21% | 0 | 0 | — |
case-02 | fail→pass | 32,266 | 22,559 | -30% | 1 | 1 | 0% | 5,111 | 3,982 | -22% | 0 | 0 | — |
case-03 | pass→pass | 21,132 | 23,400 | +11% | 1 | 1 | 0% | 2,839 | 3,260 | +15% | 0 | 0 | — |
case-04 | pass→pass | 22,878 | 20,463 | -11% | 1 | 1 | 0% | 3,269 | 3,338 | +2% | 0 | 0 | — |
case-05 | pass→pass | 54,230 | 27,322 | -50% | 1 | 1 | 0% | 8,247 | 5,453 | -34% | 0 | 0 | — |
case-06 | fail→fail | 22,287 | 13,445 | -40% | 1 | 1 | 0% | 3,193 | 2,980 | -7% | 0 | 0 | — |
case-07 | fail→fail | 34,573 | 15,366 | -56% | 1 | 1 | 0% | 4,931 | 3,428 | -30% | 0 | 0 | — |
case-08 | pass→pass | 14,212 | 17,023 | +20% | 1 | 1 | 0% | 2,121 | 2,810 | +32% | 0 | 0 | — |
case-09 | fail→pass | 9,872 | 9,665 | -2% | 1 | 1 | 0% | 1,476 | 1,944 | +32% | 0 | 0 | — |
case-10 | fail→fail | 28,985 | 22,765 | -21% | 1 | 1 | 0% | 4,561 | 5,227 | +15% | 0 | 0 | — |
case-11 | pass→fail | 8,348 | 4,816 | -42% | 1 | 1 | 0% | 1,414 | 1,419 | +0% | 0 | 0 | — |
case-12 | pass→pass | 10,933 | 11,441 | +5% | 1 | 1 | 0% | 1,752 | 2,616 | +49% | 0 | 0 | — |
case-13 | fail→pass | 15,419 | 7,386 | -52% | 1 | 1 | 0% | 2,202 | 1,494 | -32% | 0 | 0 | — |
case-14 | fail→fail | 9,553 | 10,901 | +14% | 1 | 1 | 0% | 1,500 | 2,083 | +39% | 0 | 0 | — |
case-15 | pass→pass | 19,904 | 11,904 | -40% | 1 | 1 | 0% | 3,136 | 2,911 | -7% | 0 | 0 | — |
case-16 | fail→fail | 23,799 | 21,291 | -11% | 1 | 1 | 0% | 3,163 | 3,625 | +15% | 0 | 0 | — |
case-17 | pass→pass | 12,460 | 12,109 | -3% | 1 | 1 | 0% | 2,272 | 2,722 | +20% | 0 | 0 | — |
case-18 | pass→pass | 12,872 | 12,580 | -2% | 1 | 1 | 0% | 1,716 | 1,604 | -7% | 0 | 0 | — |
case-19 | pass→pass | 9,072 | 9,268 | +2% | 1 | 1 | 0% | 1,563 | 2,290 | +47% | 0 | 0 | — |
case-20 | pass→pass | 12,915 | 9,896 | -23% | 1 | 1 | 0% | 1,907 | 2,490 | +31% | 0 | 0 | — |
case-21 | pass→pass | 12,195 | 9,740 | -20% | 1 | 1 | 0% | 2,324 | 2,293 | -1% | 0 | 0 | — |
case-22 | pass→pass | 14,196 | 10,947 | -23% | 1 | 1 | 0% | 1,840 | 2,530 | +38% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +9 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.