[WIP] Refactor llms-stack-refresh for AI-agent-first documentation

[Copilot]

• Explore repository structure and current file contents
• Delete docs/CONTRIBUTING.md
• Rewrite README.md — remove emojis, remove Contributing section, expand plugin/extension and coding tool integration sections
• Rewrite AGENTS.md — remove "Editing or adding docs", add Consumption Protocol, Context Window Strategy, Coding Tool Config Map, Stack Integration Rules
• Update docs/INDEX.md — remove CONTRIBUTING.md row, add Coding Tool Configuration section
• Update docs/PROVENANCE.md — reframe as agent integrity verification protocol, remove human contribution language
• Final code review and security scan

Original prompt

Objective

Refactor the entire llms-stack-refresh repository to be AI-agent-first only. This is a documentation-only repository containing curated llms.txt files for a full-stack TypeScript/Bun workflow (Bun, Elysia, htmx, Prisma, Tailwind CSS, daisyUI, EasyAuth). The repository needs significant improvements to make it specifically optimized for AI agent consumption.

Current Repository Structure

llms-stack-refresh/
├── README.md            ← Main entry point (has emojis, contributing references)
├── AGENTS.md            ← AI agent navigation guide (needs expansion)
├── bun/llms.txt         ← Official Bun docs
├── daisyui/llms.txt     ← Official daisyUI docs
├── easy-auth/llms.txt   ← Curated EasyAuth docs
├── elysiajs/llms.txt    ← Curated ElysiaJS docs
├── htmx/llms.txt        ← Curated htmx docs
├── prisma/llms.txt      ← Curated Prisma docs
├── tailwindcss/llms.txt ← Curated Tailwind CSS docs
└── docs/
    ├── INDEX.md         ← Navigation map
    ├── CONTRIBUTING.md  ← Contributing guide (TO BE REMOVED)
    └── PROVENANCE.md    ← Provenance policy

Changes Required

1. Remove All Contributing References and Emojis

• Delete docs/CONTRIBUTING.md entirely — this repository is AI-agent-first only, not for human open-source contributors.
• Remove all emojis from README.md (the table currently has 🍞, 🌼, 🔐, ✨, 🟦, 🔺, 🌬️ in the "Included Files" table and elsewhere).
• Remove the entire "Contributing" section from README.md (lines 152-162 starting with ## Contributing).
• Remove the CONTRIBUTING.md reference from docs/INDEX.md in the "Supporting Docs" table.
• Remove the CONTRIBUTING.md reference from the repository map in README.md and AGENTS.md.
• Remove "Editing or adding docs" section from AGENTS.md — agents don't contribute, they consume.

2. Improve README.md for AI Agent Consumption

Transform the README from a human-oriented contributing repo overview into an AI-agent-first operational reference:

• Replace the intro paragraph to clearly state this is an AI-agent-first reference collection — not a community project.
• Rewrite the "Stack Composition" section with more technical precision — include version constraints, API surface areas, and integration contract details.
• Expand the "Plugin / Extension / Customization" section significantly with actual code patterns and method signatures for each stack (see below).
• Rewrite the "Using This Repository with AI Tools" section to be specific per coding tool with detailed integration instructions mapped to each llms.txt file.

3. Massively Expand Extension/Plugin System Documentation

The current content mentions plugins/extensions superficially. Each needs concrete, actionable detail that an AI agent can use to generate correct code. Add a new top-level section or significantly expand the existing one in README.md with:

Bun Plugin System

• Bun.plugin() API signature and lifecycle
• Plugin setup() hook with onLoad, onResolve callbacks
• Loader configuration (.toml, .yaml, custom extensions)
• Macro system (with { type: "macro" } imports)
• bunfig.toml configuration reference (preload, plugins, install, test sections)
• Workspace configuration (workspaces in package.json, filter patterns)

Elysia Plugin System

• .use() method signature and plugin deduplication (name + seed)
• Plugin lifecycle hooks: onRequest, onParse, onTransform, onBeforeHandle, onAfterHandle, onAfterResponse, onError, onStop
• .decorate(), .derive(), .model(), .macro(), .guard(), .state(), .resolve()
• Official plugins with install commands: @elysiajs/bearer, @elysiajs/cors, @elysiajs/cron, @elysiajs/html, @elysiajs/jwt, @elysiajs/openapi, @elysiajs/opentelemetry, @elysiajs/server-timing, @elysiajs/static
• Eden Treaty client generation for E2E type safety

htmx Extension System

• htmx.defineExtension() API
• Extension hook signatures: init, getSelectors, onEvent, transformResponse, isInlineSwap, handleSwap, encodeParameters
• Core extensions list: head-support, loading-states, multi-swap, response-targets, ws, sse, json-enc, preload
• How to build a custom extension (skeleton code)
• hx-ext attribute usage for activating extensions

Prisma Client Extensions ($extends)

• Four extension component types: client, model, query, result
• $extends() method signature
• Computed fields via result component
• Custom methods via model component
• Query interception via query component
• Client-level methods via client component
• Shared/reusable extension patterns
• Stacking multiple extensions

Tailwind CSS Customization

• @plugin directive for writing plugins in CSS
• @utility for custom utility registration
• @variant and @custom-variant for custom variants
• @apply for extracting utility patterns
• @reference for type-safe IDE hints
• `@the...

This pull request was created from Copilot chat.

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.

[Copilot]

Copilot wasn't able to review any files in this pull request.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.