Status: Approved in brainstorming session 2026-07-07
Owner: vikas (human) — executes by dispatching AI per mini-plan using superpowers
Repo: pink-pulsar — Astro 7 + @astrojs/cloudflare adapter on Cloudflare Workers
Purpose: A plan-of-plans for a production-grade end-to-end e-commerce website. Each mini-plan is one PR-sized, AI-implementable, fully-tested increment. The plan-of-plans itself is for human use only — the AI does not execute it; it executes one mini-plan at a time on demand.
A pattern for building personal knowledge bases using LLMs.
This is an idea file, it is designed to be copy pasted to your own LLM Agent (e.g. OpenAI Codex, Claude Code, OpenCode / Pi, or etc.). Its goal is to communicate the high level idea, but your agent will build out the specifics in collaboration with you.
Most people's experience with LLMs and documents looks like RAG: you upload a collection of files, the LLM retrieves relevant chunks at query time, and generates an answer. This works, but the LLM is rediscovering knowledge from scratch on every question. There's no accumulation. Ask a subtle question that requires synthesizing five documents, and the LLM has to find and piece together the relevant fragments every time. Nothing is built up. NotebookLM, ChatGPT file uploads, and most RAG systems work this way.
- Focus on clarity: Make your message really easy to understand.
- Example: "Please send the file by Monday."
- Be direct and concise: Get to the point; remove unnecessary words.
- Example: "We should meet tomorrow."
- Use simple language: Write plainly with short sentences.
- Example: "I need help with this issue."
- Stay away from fluff: Avoid unnecessary adjectives and adverbs.
- Example: "We finished the task."
| import anthropic | |
| import os | |
| import sys | |
| from termcolor import colored | |
| from dotenv import load_dotenv | |
| class ClaudeAgent: | |
| def __init__(self, api_key=None, model="claude-3-7-sonnet-20250219", max_tokens=4000): | |
| """Initialize the Claude agent with API key and model.""" |
| import anthropic | |
| import os | |
| import sys | |
| from termcolor import colored | |
| from dotenv import load_dotenv | |
| class ClaudeAgent: | |
| def __init__(self, api_key=None, model="claude-3-7-sonnet-20250219", max_tokens=4000): | |
| """Initialize the Claude agent with API key and model.""" |
| // Update globs depending on your framework | |
| --- | |
| name: tailwind_v4 | |
| description: Guide for using Tailwind CSS v4 instead of v3.x | |
| globs: ["**/*.{js,ts,jsx,tsx,mdx,css}"] | |
| tags: | |
| - tailwind | |
| - css | |
| --- |
| import { useState, useEffect } from "react"; | |
| export const useFetch = (url, ref, initialValue) => { | |
| const [data, setData] = useState(initialValue); | |
| const [error, setError] = useState(null); | |
| const [loading, setLoading] = useState(true); | |
| useEffect(() => { | |
| if (ref.current) { | |
| (async () => { |
| import { useState, useEffect } from "react"; | |
| export const useFetch = (url, ref, initialValue) => { | |
| const [data, setData] = useState(initialValue); | |
| const [error, setError] = useState(null); | |
| const [loading, setLoading] = useState(true); | |
| useEffect(() => { | |
| if (ref.current) { | |
| (async () => { |
| import { useReducer } from 'react'; | |
| const useUndoReducer = (reducer, initialState) => { | |
| const undoState = { | |
| past: [], | |
| present: initialState, | |
| future: [] | |
| }; | |
| const undoReducer = (state, action) => { |
| /** | |
| * List unique CSS properties for all DOM elements | |
| * Initially created to list unique font stacks on a page | |
| * @see {@link http://stackoverflow.com/a/35022690/ Inspired by this StackOverflow answer} | |
| * | |
| * @see {@link https://gist.github.com/macbookandrew/f33dbbc0aa582d0515919dc5fb95c00a/ URL for this file} | |
| * | |
| * @author AndrewRMinion Design (https://andrewrminion.com) | |
| * @version 1.1 | |
| * |