| name | pmcro-project-generator |
|---|---|
| description | Generates a complete, production-ready PMCR-O Cognitive Architecture .NET project as a downloadable ZIP file. Use this skill whenever a user provides a project name and seed intent and wants to scaffold a full PMCR-O project. Triggers on phrases like: "generate PMCR-O project", "create cognitive architecture project", "scaffold PMCR-O", "new PMCR-O project", "generate project zip", or any combination of intent + project name that implies a PMCR-O / cognitive architecture scaffold is desired. Also triggers when user provides an intent string and asks for a project, template, or architecture to be generated around it. |
Discover gists
| <!DOCTYPE html> | |
| <html lang="zh-CN"> | |
| <head> | |
| <meta charset="UTF-8"> | |
| <meta name="viewport" content="width=device-width, initial-scale=1.0"> | |
| <title>2026年04月06日 AI 早报</title> | |
| <style> | |
| :root { | |
| --c-bg: #f4f5f7; | |
| --c-surface: #ffffff; |
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.
| <?php | |
| /** | |
| * EDD REST API - API Key Authentication Example | |
| * | |
| * The EDD REST API uses a public key + secret token pair. | |
| * Generate these under Users > Your Profile > EDD API Keys. | |
| * | |
| * Authentication via query string: | |
| * https://yourstore.com/edd-api/v2/products?key=PUBLIC_KEY&token=TOKEN | |
| * |
| [00:00:00] of multiple things like chaining of multiple you know sending props like sending props to child and then | |
| [00:00:11] like grandchild and then like and so on so this is probably so have you heard about user | |
| [00:00:20] Yes, sir. | |
| [00:00:21] I have also heard about you. | |
| [00:00:24] So what is the difference between use state and use ref hooks? | |
| [00:00:28] Sir, use state hook is basically used to like perform re-rendering. | |
| [00:00:33] But when we use use ref, so there is no re-rendering happens. | |
| [00:00:40] Can you give me an example of user? | |
| [00:00:43] Sir, what? | |
| [00:00:44] Can you give me any generic example where you can use user? |
| Build a React chat input component styled with shadcn/ui and Tailwind. White background, rounded-lg border, subtle shadow. Inside the box: | |
| A <textarea> with placeholder "How can I help you today?" that auto-grows as the user types (supports multiple lines). No visible border, transparent background. | |
| Below the textarea, a bottom toolbar row with: a "+" icon button on the left, and on the right side a model selector dropdown ("Opus 4.6 Extended" with a chevron-down) and a voice/audio waveform icon button. | |
| Use shadcn Button (variant ghost/icon), DropdownMenu, and Textarea. Keep everything inside a single rounded container with bg-white border border-gray-200 rounded-xl p-3 flex flex-col gap-2. The toolbar uses flex justify-between items-center. Make it responsive and max-width around 700px centered. |
| // for more than 4K locations or for slow network you need to gzip the body | |
| OkHttpClient client = new OkHttpClient(); | |
| client.setConnectTimeout(20, TimeUnit.SECONDS); | |
| client.setReadTimeout(20, TimeUnit.SECONDS); | |
| client.interceptors().add(new GzipRequestInterceptor()); | |
| .... | |
Claude Code is powerful. It can write code, run commands, push to git, and modify any file on your system.
That last part should concern you.
By default, Claude Code has no safety rails. No command is blocked. No file is protected. A user recently lost their entire C:\Users directory when Claude ran rm -rf that followed NTFS junctions.
Here's how to fix that in 60 seconds.
| Log uploaded on Monday, April 6, 2026, 1:45:32 PM | |
| Loaded mods: | |
| Hair Layer Fix(pikuSucks.HairLayerFix): HairLayerFix(1.0.0) | |
| Harmony(brrainz.harmony)[mv:2.4.2.0]: 0Harmony(2.4.1), HarmonyMod(2.4.2) | |
| Core(Ludeon.RimWorld): (no assemblies) | |
| Shuttle Dock(Salvador143.ShuttleDock): ShuttleDock(1.0.6) | |
| Load in Background(ferny.LoadInBackground): LoadInBackground(1.0.0) | |
| [CAT] Show Hair With Hats or Hide All Hats(cat2002.showhair): ShowHair(0.0.0) | |
| [NL] Custom Portraits(Nals.CustomPortraits): CustomPortraits(1.0.0) | |
| Auto links(automatic.autolinks): 0Harmony(av:2.4.1,fv:2.0.2), Autolinks(1.0.379.18) |