Two friends learning AI engineering, and the game we built to keep each other going: XP, quiz duels, code races and an in-browser AI lab.

SIMIT: Self-Improving Vision-Language Models via Imagination at Test-Time, official implementation of the paper

A memory for coding agents: pin what you were told, journal what you did, sleep when your head is full. pi extension, Claude Code plugin, Codex/Copilot hooks, and the benchmark behind it.

A self-hosted gateway that serves your Claude and ChatGPT subscriptions as OpenAI and Anthropic compatible APIs, with per-user keys, account pooling, live quota tracking and an admin dashboard.

Voice to structured notes and diagrams. Talk, and a local model builds the headings, lists, tables and diagrams while you think, then typesets a PDF. Hands-free or one-handed with an MCP toolbox for AI agents.

Floating Tree by sudoingx: a one-file prompt benchmark. Hand FLOATING_TREE.md to any coding agent and it builds a 3D floating island in Three.js, graded by 12 tests and a visual gate.

🫕 One pot, everyone brings a fork, no double-dipping. A team of AI agents that plan, build and review each other's work, with you as the team lead. Spec-driven development and bug hunting for Claude Code.

jaja: a job-search agency for Claude Code. Agents find roles and people, tailor pages from your evidence and never send without your yes.

pxchat plus 增强 | nonebot-plugin-pxchat 的增强补丁：修复稳定性问题（调用超时/上下文污染/模型过滤自动切换/长文本处理/并发加速）。关键词：pxchat plus nonebot插件 聊天插件 增强补丁。

Croct Agent Skills for Claude Code, Cursor, and Codex: teach AI coding agents to add personalization, A/B testing, and headless CMS content with Croct.

A minimal LLM pipeline that runs on a laptop, written from scratch in PyTorch and to be read: tokenizer, pretraining, inference, fine-tuning, RLHF and RLVR, with measured results and an interactive website.

LLM-agnostic AI guardrail API with Jev. Checks prompts and responses against YAML policies and returns PASSED, BLOCKED or FAILED. Sits between your app and any LLM provider.

Self-improving skills and memory for any AI agent framework, with governance built in

Auto-route each Claude Code session to Haiku, Sonnet, Opus or Fable from its first prompt. Classified locally, zero tokens.

Persistent, embeddings-backed memory for agents. Store facts and entities, recall them semantically, and surface a working context — over the three.ws Memory engine.

把 AI 接进 QQ：NapCat + FastAPI 桥接，带多角色卡、长期记忆、情绪与生活线、识图、语音转写和 QQ 空间能力。

个人文章写作 + 去 AI 味的可复现方法论：机制与数据分离、证据优先、分级去 AI 味(L1-L6) + 语气硬门禁 | Reproducible methodology for writing in your own voice & removing AI flavor

Could AI replace you? No, actually. A private, local-first scorecard that counts the times you caught your AI being wrong. Claude Code, Codex, ChatGPT and more.

Local job search assistant for Germany: finds jobs daily, scores them against your CV, writes truthful tailored CVs and cover letters.

Shared, brain-inspired memory for AI agents: 13 brain regions with awake/sleep loops, a live visualizer, agents connected via REST/SDK/MCP, governance & audit, and business-value metrics. Works with Claude, GPT, Gemini, DeepSeek or local LLMs. SQLite storage. VI/EN/JA.

Python library for running LTspice from code and from language models: headless simulation, .meas results and short waveform summaries as JSON, and 10 tool schemas (OpenAI, Anthropic, Gemini) so a model can build, simulate and tune circuits. Tested on Windows with LTspice 26.1.1.

Describe a part in words → parametric, editable Onshape feature tree. LLM + FeatureScript + local OpenCascade builder, built for Onshape's API limits.

A Claude Code plugin that trims what every request re-sends. Same work, same quality, about half the tokens.

Stop vibe coding. Start from here. One Agent Skill for any AI. Interview until complete project blueprint (PRD, TRD, architecture, ERD, security, tests, legal, AGENTS.md) before the first line of code.

One channel for your coding-agent sessions and you

Free Cursor rules + AGENTS.md that stop AI (Cursor, Claude Code, ChatGPT) from writing broken Unity C#: missing usings, invented APIs, Unity 6 vs 2022 mixups. MIT.

Transform your FastAPI routes to agent-friendly, context-efficient MCP

Native Java Model Context Protocol (MCP) server for live Spring Boot inspection via Attach API.

How many watts does your local LLM really burn? Live power, temperature and energy-per-token dashboard for Windows: llama.cpp, Ollama, LM Studio. No admin, no drivers, 0 dependencies.

Local-only QQ group chat digester: decrypt your own nt_msg.db, build a rolling index, generate reports, and recognize key images. 本地解密 QQ 聊天库并总结群聊，不登录、不注入、不上传。