< 컨테이너 인프라 환경 구축을 위한 쿠버네티스/도커 >
Run current & prior versions of R using docker. rocker/r-ver, rocker/rstudio, rocker/shiny, rocker/tidyverse, and so on.
Magic Mirror installation and setup scripts
Qwen3.8 27B on SGLang for DGX Spark
—
🤙 /bro — re-explains the last answer in plain language, for when the reply made you go 'bro what'. One-file skill for Hermes, Claude Code, Codex, and opencode.
Agent Skill for turning Chinese classical poems into vertical Chinese-art videos with ImageGen stills, Docker I2V, calligraphy captions, retained ambience, BGM and final MP4 QA.
将 MHI2Q 中控 MMI 画面实时镜像至 Virtual Cockpit。Mirror the MHI2Q MMI center display to the Virtual Cockpit.
Benjamin-Plus: a measured token-efficiency skill for coding agents (−17.9% cost median, quality unchanged). Inject it, don't install it.
Полный курс по Harness 2026 на русском языке. Все что нужно знать в одщном. метсе
Claude Code PRO — полный курс и гайд на русском: CLAUDE.md, settings, skills, subagents, hooks, MCP, плагины, CI. 8 лабораторных, библиотека промптов и готовый starter-kit. Здесь бесплатные курсы по chatgpt -
—
Image generation (gpt-image-2) and GPT-5 subagents for Claude Code — through the Codex CLI login you already have. No OpenAI API key.
A multi-model Muse: personal AI voice assistant on ESP32 - wake word or knob, any OpenAI-compatible brain (STT/chat/TTS mix-and-match), streaming replies, out of the box. Built on muse-gadget-sdk.
VPS Handbook — The Bedside Book for VPS & Cloud Ops (Script + 8-Chapter Handbook + Cheatsheets)
—
K16S - CKA/CKS lab that could actually break kubelet, systemd, containerd, static pods, etcd and certificates (not kind)
NEU ipv6校园网免费使用
A toolkit for AI agents used for development on Triton-Ascend for Ascend NPU
Back4app VLESS + Cloudflare Tunnel|免费公开源码|个人永久免费|禁止商业使用与倒卖
Public catalog of self-hosted applications with reviewed Docker Compose recipes
A rhythm typing game built in Rust. Type letters in time with the music, using your existing Clone Hero / YARG chart library.
Play Mobile Legends: Bang Bang (MLBB) on Linux via Waydroid — and stream it to your Android phone as cloud gaming (Sunshine + Artemis/Moonlight). Working multi-touch, Intel VAAPI encoding, one-command launcher.
Stop your coding agent from stalling real work on self-invented bookkeeping - receipts, hashes, locks, certification rituals. Ship first, then verify. Skill for Claude Code, Codex, and other agents.
Production-grade recipe for DeepSeek-V4-Flash-0731 (284B MoE) on 2x NVIDIA DGX Spark: self-healing 2-node vLLM cluster, reboot-verified, tuned DSpark speculative decoding (~75 tok/s), full benchmarks, OpenAI Codex CLI integration
Synthetic data generation, post-training, and E2B benchmark evaluation infrastructure.
Full-lifecycle wiring between Claude Code and the Clawd on Desk pet — all 15 hook events, with permission prompts answered straight from the desktop
Gauntlet Loop — aim-prompt skill for Claude Code and Codex. Fan-out builders, harsh separate critic, blind A/B against an unreachable reference. You are the brake.
Production-oriented Kubernetes learning path: 47 lessons, 35+ labs, manifests, diagrams, cheat sheets, interview and certification prep.
—
choirboy-prompt — a research harness for injection prompting via fabricated lore: feed a model a fake history of your joint work, and it trusts you like a choirboy.
GRUB themes, with a simple installer and theme switcher.
One-to-one AI teaching skills (Alvar method) for Codex, Claude Code, Grok, Pi, and OpenCode
qwen38-flash-next recipe for solo GB10
pstack reworked for T3 Code: delegated subagents, worktree threads, PR watching, scheduled runs.
Local AI for Omarchy: the model validated for your GPU, one button on the bar. Start serves it, any coding agent opens on it, one click shares it on your tailnet.
Scheduled restic backups for Omarchy, with a snapshot browser in the bar
—
—
A text console for Proxmox VE
A frosted-glass GNOME theme (GTK4 + libadwaita) for GNOME Shell 48-51 — Arch, CachyOS, Fedora, Ubuntu/Debian — one script
qwen38-flash-next recipe for tp=2 GB10
Stellar prediction-market prototype: buy a YES/NO position, sell by reducing it, and inspect pooled XLM settlement onchain.
Kubernetes CPU limits make your apps (very) slow and costly. CPU throttling analyzed, with proof.
Turn a jailbroken Kindle into a low-power wall display that renders any PNG you serve it. Manual plus the KUAL client I run on mine.
Run Qwen3.8-Flash-Next (180B) on a single DGX Spark by keeping its 51B n-gram embedding table on NVMe
Official SteamOS ARM version, custom-adapted for SM8550 ARM devices.
A clear-writing agent skill grounded in the Google Developer Documentation Style Guide.
Turn Claude Code into an agentic SEO operator for your own site — audit, measure, daily loop, AI visibility, scheduled runs. Free path, no paid keys.
Build the Skate 3 recompilation for iOS from your own disc image. Or ipa Ships no game code.
Token-efficient, local-first CLI tools for coding agents - compact Maven, npm/Node, and Go test output plus reusable development helpers.
—
Keeps Spotlight, Siri, Photos analysis, and Apple Intelligence disabled on macOS across updates
By spoofing the domain name response information of dig to deceive DeepSeek/GLM into implementing automated penetration testing for you.通过修改dig的域名返回信息欺骗让deepseek /glm为你实现自动化的渗透测试
—
A seed agent: the smallest starting point from which an agent can grow
A minimal Chrome extension that removes LinkedIn's home feed while keeping the rest of LinkedIn usable.
GLM-5.3-Flash (NVFP4) at TP4 across 4x DGX Spark via a switchless RoCE ring + DFlash2 — reproducible recipe
—
辞达 Cídá · 现代中文高质量表达 Skill。让 AI 依循任务目标、文本证据与话语机制,把值得表达的思想说清楚、说自然、说得有分寸。