write-readme

>-

wilbeibi/wilbeibi-skills41 installsMITSynced Aug 26

Works with

Claude CodeCursorCodex CLIGitHub CopilotGemini CLI
---
name: write-readme
description: >-
license: MIT
---

# write-readme

Write READMEs for readers with short attention: what is it, when do I use it, how do I start, what should I not expect.

## Workflow

1. Gather context: problem solved, target audience, current alternatives, sharpest value/proof, and non-goals.
2. Classify once: library, CLI/tool, or agent tool. Ask if unclear.
3. Cut before adding: remove duplicated inventories, badge noise, marketing claims, and copied `--help` tables.
4. Draft in inverted-pyramid order: plain value, install, common use, boundaries, then optional reference.
5. Put the most common invocation or import path in the first 30 lines.
6. Use [REFERENCE.md](REFERENCE.md) only for section choices, not as a required template.
7. Final pass: every line must answer a likely reader question or change their next action.

## Opening Rules

- Lead with concrete value: metric, named capability, or sharp differentiator.
- Never start with "A library/tool that..."; show the reader's outcome first.
- Name the target audience and the non-goal early.
- Claims need evidence or a methodology note.
- Be genuine. Avoid ad-copy words like "seamless", "instant", "powerful", and "no re-explaining" unless the README proves them.
- Prefer a normal-user voice over a product-sheet voice. A first-person note is fine when it names a real daily use; cut it if it turns into taste-making, hype, or a joke.
- Open on the human moment the tool fixes, then name the mechanism. "I was here, this helped" beats "this solution enables..."

## Ruthless Cut Pass

Delete or compress:

- Repeated supported-platform or integration lists. Name them once, where they affect command syntax or installation.
- Full flag/reference tables when `--help`, generated docs, or package docs already cover them.
- Badges that do not affect trust or installation decisions.
- Separate "What it does", "Features", and "Why" sections that restate the same sentence.
- Long caveats. Keep boundaries, but write them as short operational facts.

## Agent Tools

For CLIs meant to be invoked by coding agents:

- Treat the README as the tool's function signature.
- Show exact command shape, required args, important flags, and expected success/error shape when that output helps the agent verify work.
- Add `Not for X` guardrails so agents avoid misuse.
- Keep the core interface scannable without expanding `<details>`.

## Non-Negotiables

- Install/quickstart before architecture.
- Limitations get their own heading, not a buried caveat.
- Broad support claims must name platforms, formats, languages, or integrations.
- Troubleshooting belongs in the README only when the failure is common and actionable.
- Use `<details>` only for advanced config, raw data, methodology, or alternative setup.
- No marketing adjectives; use numbers, examples, and named evidence.
- No dramatic connector cadence as a default voice: em dashes, arrows, and "not just X, but Y" need a reason.

See [REFERENCE.md](REFERENCE.md) for templates, badge patterns, screenshot guidance, tone notes, and final-pass prompts.

More Writing & Documentation skills

paper-context-resolver

lllllllama/rigorpilot-skills

Rigor Paper Context helper for README-first deep learning repo reproduction. Use only when the README and repository files leave a narrow reproduction-critical gap and the task is to resolve a specific paper detail such as dataset split, preprocessing, evaluation protocol, checkpoint mapping, or runtime assumption from primary paper sources while recording conflicts. Do not use for general paper summary, repo scanning, environment setup, command execution, title-only paper lookup, or replacing README guidance by default.

450.8k

repo-intake-and-plan

lllllllama/rigorpilot-skills

Rigor Intake helper for README-first deep learning repo reproduction. Use when the task is specifically to scan a repository, read the README and common project files, extract documented commands, classify inference, evaluation, and training candidates, and return the smallest trustworthy reproduction plan to the main orchestrator. Do not use for environment setup, asset download, command execution, final reporting, paper lookup, or end-to-end orchestration.

450.0k

minimal-run-and-audit

lllllllama/rigorpilot-skills

Rigor Run skill for README-first deep learning repo reproduction. Use when the task is specifically to capture or normalize evidence from the selected smoke test or documented inference or evaluation command and write standardized `repro_outputs/` files, including patch notes when repository files changed. Do not use for training execution, initial repo intake, generic environment setup, paper lookup, target selection, hidden scientific-meaning changes, or end-to-end orchestration by itself.

449.9k

← All Writing & Documentation skills

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