agentleFS
TylerAkinsTylerAkins/ai-agent-skills/skills/readme-design/SKILL.md

readme-design

Skill1 starsChanged 32 days agoMITView on GitHub

Design, rewrite, or review a project's README as its public entrypoint. Use when readers need to understand the project, choose a path, get started, and find deeper documentation. Use technical-documentation for source-derived guides and references.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

9 rules. About 0 are common, 9 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈9

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read anti-ai-writing before drafting.

    line 12
  2. Distinctive:

    Do not infer behavior from filenames or dependency presence.

    line 23
  3. Distinctive:

    Start with a clear title and a short description of what the project does and who it helps.

    line 28
  4. Distinctive:

    Keep shared prerequisites and navigation visible.

    line 53
  5. Distinctive:

    Use GFM for Markdown READMEs unless the destination explicitly requires another format.

    line 68
  6. Distinctive:

    Use a logical heading hierarchy, focused paragraphs, descriptive links, inline code for identifiers, and language-tagged code fences.

    line 73
  7. Distinctive:

    Avoid wide tables, nested disclosures, decorative separators, and repeated headings.

    line 76
  8. Distinctive:

    Use a short navigation row or table of contents when it helps readers find their path.

    line 77
  9. Distinctive:

    Check local links, heading anchors, referenced files, and commands.

    line 83

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

README Design

Read anti-ai-writing before drafting. Make the README human-readable, visually coherent, and useful from first glance to first use. Preserve project identity and accurate existing content. Review-only requests do not authorize edits; README work does not authorize publication.

1. Establish the reader and facts

Inspect repository instructions, the current README, linked docs, and relevant source, manifests, configuration, and scripts. Establish the intended audience, project purpose, supported capabilities, and paths to first use. Trace commands, prerequisites, and feature claims to implementation or authoritative documentation. Do not infer behavior from filenames or dependency presence. Resolve missing facts before publishing dependent claims, or omit them and report the gap.

2. Design the opening and reader path

Start with a clear title and a short description of what the project does and who it helps. Use existing branding where useful. Add a screenshot, demo, diagram, or representative output when it makes the result easier to understand. Do not force a banner or visual onto a project that needs only a small working example.

Choose a few useful, current badges with consistent styling and meaningful link targets. Omit vanity metrics, redundant signals, and badges that overwhelm the introduction. Check visual assets for readability on light and dark backgrounds, provide useful alt text, and avoid essential information available only in images.

Arrange content so readers can understand the project, decide whether it fits, choose a supported path, reach a first useful result, and find deeper documentation. Show commands in context with prerequisites and an expected outcome. Make install choices and first use easy to find without a long preamble. When presentation is weak or a redesign is requested, consult design examples for specific patterns rather than copying another project's layout wholesale.

3. Keep every section useful

Choose sections for this project's reader needs. Do not fill a fixed template. Features, architecture, FAQ, roadmap, acknowledgments, contribution, and license sections belong only when they add supported information the reader needs. Link to existing repository files rather than repeating their contents. Remove unused placeholders, repetitive summaries, and inventories that do not help first use.

Keep shared prerequisites and navigation visible. Use clearly labeled collapsible sections for independent installation paths or optional detail, with each expanded path complete enough to follow. Preserve all supported alternatives and required install, activation, and update instructions; brevity does not justify deleting paths. Keep platform-specific commands clearly labeled and treat supported tools equally unless the project has a justified recommendation. Use generic examples unless a named tool is necessary for the actual project.

Link long troubleshooting, exhaustive configuration, and maintainer procedures to focused documents. Use technical-documentation for those guides. Follow repository requirements when procedures must remain in the README; do not replace required instructions with a link.

4. Format with GitHub Flavored Markdown

Use GFM for Markdown READMEs unless the destination explicitly requires another format. Apply GFM guidance when selecting or checking syntax. Distinguish GFM from GitHub-specific features and check support on the target surface.

Use a logical heading hierarchy, focused paragraphs, descriptive links, inline code for identifiers, and language-tagged code fences. Use tables for compact comparisons, ordered lists for steps, and task lists only for actual tracked work. Avoid wide tables, nested disclosures, decorative separators, and repeated headings. Use a short navigation row or table of contents when it helps readers find their path. Keep spacing and visual styling consistent; add formatting only when it improves comprehension or navigation.

5. Verify the result

Check local links, heading anchors, referenced files, and commands. Follow one representative first-use path from the opening to the expected result, using safe commands where feasible. Check every required installation alternative for completeness. Change generated content at its maintained source; claim generation only when a verified generator exists.

Inspect the rendered destination when available: opening hierarchy, spacing, navigation, code, tables, images, and collapsed sections in both states. Check narrow layouts and light/dark readability when relevant. If rendering or commands cannot be checked, report the specific limit and perform available structural checks. Markdown parsing does not prove usability or working commands.

Deliver the changed artifact, checks actually performed, and material limits.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.