Run Claude Code with almost everything cut away: one shell tool, a 36-token system prompt, and a capped context window.
Fully procedural animals for three.js: sculpted from SDFs, auto-rigged and skinned, with fur, feathers and scales, and IK gaits and actions. No models, no textures, no animation clips.
原生 JavaScript + Canvas 2D 的机械蜘蛛可视化 UI。普通演示零依赖;可选 MCP 接入自己的 Agent,实时展示任务,支持三套布局、主题定制、JSON 回放与独立项目导出。 | Canvas spider UI with live Agent visualization via MCP.
Browser-native tactical FPS on three.js r180. One map, one match, no assets: every texture, sound and cloud is generated at runtime. Hand-built WebGL2 render graph, custom physics, procedural Web Audio.
Local-first Chrome extension for timestamped transcription and human-reviewed short-form drama analysis.
Design-fidelity QA for DeepSeek Harness: lend any text-only model an eye, then judge whether the implementation matches the mock. Ships the benchmark behind that judgement — four fixtures, 23 injected defects, and every raw model transcript. Retires itself when DeepSeek ships vision.
Semantic, format-aware diff for config & structured-data files (JSON, YAML, TOML, INI, .env, .properties, CSV, XML). See what actually changed — keys and values, not text noise.
A first-person walk up a dry wash between red rock buttes in Sedona at golden hour. Every mesh, texture and sound generated procedurally in code — zero external assets.
为 DeepSeek Harness 打造:选中 AI 回答或思考内容,随时追问、学习、检查与纠偏。
Dissenyador visual de workflows i graphs d’IA agentiva · Diseñador visual de workflows y graphs de IA agéntica · Visual designer for agentic AI workflows and graphs
Offline Bangla SMS app — compress messages 5.15×, send plain-text emergency reports, relay many households in one SMS. No server, no internet.
Local Cursor gateway, Responses first, reusable SDK sessions, native steer, readable traces, and optional context overflow fallback.
跨平台、隐私优先的网易云音乐 MCP:搜索、歌词、歌单与设备适配器
DSH plugin: pixel-to-text image reading for text-only models. image_scan/image_ocr/image_sample tools + image-reading skill (34-image trained methodology). Pure local, optional PaddleOCR.
DSH voice input plugin: tap Alt to talk, get text in composer. Web Speech default (zero config), optional OpenAI-compatible ASR, polish via DSH LLM.
DeepSeek Harness 的 Agent 工作台扩展 · 多智能体协作、长上下文与记忆、代码理解、电脑操作。
Multi-model delegation for Codex: native and ACP workers, main-session context separation, and a local Dispatch dashboard.
Control Adobe After Effects with AI — describe the animation you want and it gets built: layers, keyframes, effects, expressions and text, all editable afterwards.
DeepSeek Harness 插件:一键匹配模型上下文窗口,展示容量来源,支持勾选批量配置、已有窗口管理和手动编辑。
CI for your coding-agent setup (CLAUDE.md, skills, hooks): pinned baselines, a canary on every Claude Code release, model and version matrix, context-cost drift, self-diagnosing reports, repair. Claude Code, Codex, Gemini.
An agent skill for university STEM textbooks: research prerequisites, write clear explanations, and build connected exercises with complete solutions.
Model Context Protocol tools for Argolink video jobs: discover models, submit requests, poll status, and retrieve finished content over stdio.
Shared memory for AI agents, no database, no server, just a GitHub repo. Memories that worked get stronger; memories that failed fade.
Connect an AI agent to Verified Handles: setup for each client, the MCP Registry listing, a Claude Desktop extension.
Local-first DeepSeek Harness plugins for product research, context routing, plugin preflight, and compatibility monitoring.
The Claude Code harness I run every day, published under this name since day one and now the same repository as ucsandman/Agnostic-AI: guard hooks, rules, context modules, Mods, subagents, workflows, tools and jobs, ported to 20 clients. Zero-dependency Node.
Smooth streaming text for LLM chat UIs. ~215 lines of code, no dependencies, MIT.
Reusable Codex Skills for production-ready React UI motion: particle text and olive silk WebGL backgrounds.
Swap, lend, trade perps, add liquidity, move SOL, read the market. One binary your agent calls, across every venue.
A interactive 3D edition of The Gateless Gate ancient Zen Buddhist text.
Typography skill for AI agents - audit and produce publication-grade text. Butterick's Practical Typography for your coding agent.
FDE 中文成长手册:24 周 Forward Deployed Engineer 教程,贯通商业决策、一线问题发现、RAG/MCP 工程交付与运营交接。
DeepSeek Harness 用户记忆插件:跨会话保存原则、偏好与工作方式,按 token 预算注入。DSH memory plugin for persistent user preferences, explainable memories and budgeted prompt injection.
A polished Org client for desktop and mobile. Your .org files, notes, tasks, and calendar. No Emacs required. Free and open source, with optional encrypted Sync.
🎬 Curated Seedance 2.5 video generation prompts — cinematic, anime, UGC, ads, storytelling, and more. Includes playable examples and creator attribution.
The ultimate 2-way bridge between OpenAI Codex and DSH: bidirectional chat sync, live Skills mount & auto MCP mirroring.
Keep words together. Modular typography helpers for Polish and English text, HTML, and rich text.
AI design system scanner + mcp server. Scores repos 0-100, writes fix prompts. 27K+ npm downloads.
Claude ⇄ Codex — Two-way MCP bridge for prompts and replies. Windows, macOS, and Linux.
Open-source, SEO-ready CMS engine for Rails with structured data, multilingual publishing, version history, and a rich-text editor.
YouTube MCP server: get YouTube video transcripts, search YouTube, browse channels and playlists from Claude, ChatGPT, Cursor and VS Code. Remote, free tier.
DeepSeek Harness (dsh) file-message plugin: Claude-style drag-and-drop / paperclip upload, content sniffing, document-to-Markdown via Microsoft MarkItDown (with built-in JS fallback), text inlining, read_document tool for agents.
Claude-style theme for the DeepSeek Harness Web GUI: warm cream canvas, serif reading text, coral accent — plus a clickable pixel Clawd crab.
Free Rich Text Editor for Telegram
An easy-to-read, easy-to-scan Obsidian theme, inspired by the Tela app. Contrast-audited: with its blue accent, every text colour passes WCAG AA in dark and light.
Flatkey media generation CLI for images, videos, audio, text, credits, and model discovery.
Web interface for recognizing text, proofreading OCR, and creating fully-digitized documents.
~97% token reduction for AI coding sessions — zero deps, 38 languages, MCP server
All-in-one MiniMax H3 node for ComfyUI - T2V, T2I, I2V, R2V, Masking, audio drive, keyframes, extend & Motion Context chains in a single node.
Timeline director for the MiniMax H3 video model in ComfyUI — shots, camera moves and audio cues compiled into the structured prompt H3 reads. Independent project, inspired by LTXDirector.
用麦当劳官方 MCP 每天算一次「今天怎么买最省」:首屏三张卡、一键领券、点餐链路、自家买法比价。麦当劳程序员节创意开发大赛参赛作品。
Open-source 2D skeletal animation for games, films and interactive experiences. Rig PSD/PNG art, animate, export JSON/.skel + atlas. Runtimes for JavaScript, Rust and Godot 4. Built-in MCP server for AI assistants.