JSPM

fable5-method

1.0.3
    • ESM via JSPM
    • ES Module Entrypoint
    • Export Map
    • Keywords
    • License
    • Repository URL
    • TypeScript Types
    • README
    • Created
    • Published
    • Downloads 34
    • Score
      100M100P100Q100004F
    • License MIT

    Install the fable5-method skill — a model-agnostic operating discipline that makes any AI coding assistant work like a senior engineer (diagnose from real evidence, prove by running, deterministic guarantees, adversarial self-review, respect prior decisions, leave it navigable).

    Package Exports

      This package does not declare an exports field, so the exports above have been automatically detected and optimized by JSPM instead. If any package subpath is missing, it is recommended to post an issue to the original package (fable5-method) to support the "exports" field. If that is not possible, create a JSPM override to customize the exports field for this package.

      Readme

      fable5-method

      Turn any AI coding assistant into a careful senior engineer.

      It checks real evidence before changing code, proves things actually work, never breaks what you didn't ask for, and leaves your project clean. Works with Claude, ChatGPT, Cursor, Gemini, and Copilot.

      npm version license works with

      🇮🇱 מדריך התקנה בעברית →


      Install

      Pick one — you don't need both.

      ▸ Claude Code — in VS Code, Cursor, or any terminal

      npx fable5-method

      Then restart Claude Code. It switches on automatically whenever you build or fix something. (No npx? Install Node.js once from nodejs.org.)

      ▸ Claude on the web or desktop app

      ⬇ Download fable5-method.zip — then in Claude open Settings → Skills → Add skill → Upload and choose the file.

      Using another AI? (ChatGPT, Cursor, Gemini, Copilot) — after installing, paste system-prompt.txt into the tool's custom-instructions / rules.


      What changes

      A normal AI assistant With fable5-method
      Guesses the cause of a bug Traces the real code and proves it
      Says "done" without testing Builds, runs, and checks it really works
      Quietly changes things you didn't ask for Flags it and asks first
      Over-claims it works Tells you honestly what's proven vs assumed

      Built from high-output Claude Fable 5 sessions — it's a method, not a model trait, so it lifts whatever AI you run it on.

      MIT © Amit Litman