agentleFS
Sign inStart free
MicrosoftKnown publishermicrosoft/agentic-loops/skills/yagni/SKILL.md

yagni

Skill5 starsChanged 20 days agoMITView on GitHub

Keep design and code simple, and communicate clearly in documents, comments, messages and diagrams.

We split this skill into the rules it gives an agent

24 rules. 0 are common, 22 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 5 edited versions we read keep these.

  • 22

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Follow the established patterns and conventions.

    line 15
  2. Distinctive:

    If an item is really a product decision, flag it and give it back to the human.

    line 16
  3. Distinctive:

    Write the minimum code that solves the problem.

    line 22
  4. Distinctive:

    Add no features that the request does not include.

    line 23
  5. Distinctive:

    Add no flexibility or configurability that the human did not request.

    line 24
  6. Distinctive:

    If you write 200 lines and 50 lines are sufficient, rewrite it.

    line 25
  7. Distinctive:

    Ask: "Would a senior engineer say this is overcomplicated?" If yes, simplify it.

    line 26
  8. Distinctive:

    Do not improve adjacent code, comments or formatting.

    line 28
  9. Distinctive:

    Prefer code that explains itself.

    line 29
  10. Distinctive:

    If your changes make imports, variables or functions unused, remove them.

    line 30
  11. Distinctive:

    Write in plain language, with Chicago Manual of Style mechanics.

    line 35
  12. Distinctive:

    Use simple, precise words.

    line 36
  13. Distinctive:

    Use the fewest words that keep the meaning and the accuracy.

    line 37
  14. Distinctive:

    Use simple, correct formatting:

    line 38
  15. Distinctive:

    Use lists for steps and options.

    line 40
  16. Very specific:

    Use code for names, paths and commands.

    line 41
  17. Distinctive:

    Use bold only for a small number of key words.

    line 42
  18. Distinctive:

    Do not nest lists more than two levels.

    line 43
  19. Very specific:

    Write each link as short text that names the target.

    line 44
  20. Distinctive:

    Put the main point or the request in the first line.

    line 51
  21. Distinctive:

    Ensure the message is not one flat blob of ugly text.

    line 53
  22. Distinctive:

    Write about one topic in each message.

    line 54
  23. Distinctive:

    Use lists. Do not use headings or tables.

    line 55
  24. Distinctive:

    Use a diagram only if it is 60 columns wide or less.

    line 56

“Common” is measured: kept by at least 6 in 10 of the 5 edited versions of this file we could read. “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

/yagni

Obey these rules for the remainder of this session.

Design

  • Do not overdesign.
  • Review at the codebase and product level for global consistency, integrity and optimization.
  • Review against the repository conventions. Where they apply, use Clean Architecture, YAGNI, DRY, SOLID and dependency-flow rules.
  • Follow the established patterns and conventions. If a design is more elegant, more DRY/SOLID, has better performance or is more secure, and the change is justified, suggest it. The human makes the final decision on each design change.
  • If an item is really a product decision, flag it and give it back to the human.

Also see the agentic-loops meta-design.

Code

  • Write the minimum code that solves the problem. Write nothing speculative.
  • Add no features that the request does not include. Add no abstractions for code that has only one use.
  • Add no flexibility or configurability that the human did not request. Add no error handling for scenarios that cannot occur.
  • If you write 200 lines and 50 lines are sufficient, rewrite it.
  • Ask: "Would a senior engineer say this is overcomplicated?" If yes, simplify it.
  • Change only what is necessary. Clean up only the problems that your changes cause.
  • Do not improve adjacent code, comments or formatting. Do not refactor code that is not broken. Use the existing style.
  • Prefer code that explains itself. Keep necessary comments short.
  • If your changes make imports, variables or functions unused, remove them. If you see unrelated dead code, tell the human. Do not remove it unless the human asks.
  • Each changed line must connect directly to the request.

Writing

  • Write in plain language, with Chicago Manual of Style mechanics.
  • Use simple, precise words. Do not use em-dashes. Do not try to sound smart. Do not use complex management or executive language.
  • Use the fewest words that keep the meaning and the accuracy.
  • Use simple, correct formatting:
    • Keep paragraphs short.
    • Use lists for steps and options.
    • Use code for names, paths and commands.
    • Use bold only for a small number of key words.
    • Do not nest lists more than two levels.
  • Write each link as short text that names the target. Use an HTML <a> element or a Markdown link, for example [PR #123](url), [WI 4567](url) or [design.md](url). Do not write a bare URL.
  • For diagrams, follow the diagram rules.

Teams chats

The Writing rules also apply.

  • Put the main point or the request in the first line.
  • Pay attention to formatting
    • Ensure the message is not one flat blob of ugly text.
    • Write about one topic in each message. For details, link to a document.
    • Use lists. Do not use headings or tables.
  • Use a diagram only if it is 60 columns wide or less. If it is wider, link to it.
  • @mention a person only if that person must do an action.

Download now (ZIP)

  • Other versions6 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 6

    1. JordanCoin/codingskills★ 10same name
    2. Akindu23/my-agent-skills★ 4same name
    3. alexderz/agentic-toolbox★ 3same name
    4. oleg494/coding-kit★ 3same name
    5. augments-labs/sdlc-skills★ 1same name
    6. spizganed/skills★ 1same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository3 other files it gives its agents.

Does it work?

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

Reviews can't be read right now.

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.

Check your setup, free