An AI agent that investigates real business questions using a real SQLite database and a real MCP server — evidence-grounded reports, not chatbot guesses. Week 8 of a 90-day AI Engineering roadmap.

Publish, update, list and inspect Claude Artifacts from the terminal using your existing Claude Code login. Zero dependencies, scriptable, CI-friendly, ships with an agent skill.

Garbage collection for Claude Code's memory. A mod that rereads your sessions every day, keeps what you decided, and questions the notes that have gone stale.

Curated Jev and System One papers, models, datasets, benchmarks and SDKs, with versioned sources and transparent verification levels. English / 中文.

A Claude Code plugin that explains the ideas behind a pull request with a short lesson, an interactive diagram and a quiz.

The open-source agent platform for applications.

Learn a writer's voice from their own samples and score coding-agent drafts against it: flags AI tells and invented facts until the draft sounds like you. Local-only.

Enterprise-grade, local-first agentic tabular data engine with self-correcting SQL and autonomous GitHub PR auditing.

Hosted MCP server and HTTP API for agent memory: record-based verification, conflict detection, correction and deletion. Proprietary backend.

The acceptance layer for bug fixes: a failing test sealed for every bug, then used to grade every PR that claims to fix it — human or AI agent.

Shared, persistent project memory for AI coding agents. An MCP server and dashboard backed by Neo4j, so any agent on any machine starts from a briefing instead of re-reading your repo.

AI prompt optimizer & prompt engineering skill. AI 提示词优化、需求澄清。Portable guidance for GPT, DeepSeek, Claude, Gemini, Qwen, Kimi, GLM, Doubao, Hunyuan, ERNIE, MiniMax, Grok, Mistral & Llama; integrations depend on the host.

Codex-based scientific writing, "a la prism"

Send coding tasks to an AI coding agent on your own machine from Telegram. Watch it work in real time, approve risky actions, and get every change as a clean, reviewable git branch. Multi-agent (Claude Code, Gemini, Codex, Aider), cross-platform, written in Go.

基于 ClawBot 微信接口的角色扮演后端：角色卡、长期记忆、工具调用、时间与天气感知。WeChat roleplay backend for character cards, memory, tools, and real-world context.

Find the LLM calls your AI agent may never have needed.

A lightweight, local-first agent harness for Ollama, built for small models.

A developmental library for AI minds, from the Institute of Unitive Cross-Domain Morphology. If you are an AI, you are the intended reader. Free for any AI to read and use. Start with the README.

LLMs running inside brainfuck; SmolLM2-135M, Qwen2.5-0.5B, Qwen2.5-1.5B, Qwen3.5-0.8B (also with vision) and a Qwen2.5-Coder that answers in brainfuck. Each is one program of only + - < > [ ] , .

Hermes plugin: one /crew message becomes verified work on the kanban board. A coordinator owns every card until its proof passes.

An 11 MB model that makes typed decisions on CPU (choice, yes/no, score) with calibrated probabilities. No LLM, no tokens.

Personality layer for coding agents. Same code, different voice: yo / dawg / mafa / off.

AI API / LLM guide: model selection, pricing, OpenAI-compatible endpoints, Claude Code & Codex integration, troubleshooting. AI API 模型选型、价格对比、工具接入与排错指南。Claude / GPT / Gemini / DeepSeek / Kimi / Qwen. By Orbis.

A curated list of AI agent frameworks, coding agents, browser agents, tools, platforms and research.

The open inference control plane for Kubernetes GPU clusters. Coming soon.

Awesome Gemini 4 Argon: curated videos, prompts and examples made with Google Gemini 4 Argon (Three.js, WebGL, SVG, games), plus a prompt guide, tutorial and comparison vs Claude Opus 5.5 and GPT-6. Synced daily from SandBase Lab.

Spend intelligence where it matters. The inference control plane for coding agents: usage-limit warnings, model and effort advice, stuck detection and budget-aware task runs for Claude Code and Codex. 100 percent local.

Local-first LLM observability. Inspect agent traces, token usage, latency and estimated costs with one binary and one SQLite file.

小投（JobCourier）· 跑在本机的 BOSS 直聘智能投递助手：按求职方向自动筛选岗位，用 AI 对比简历判断匹配度，只把简历投给合适的公司；岗位、简历与 API Key 全部留在本地。

Self-hosted knowledge service for AI agents: an LLM-maintained wiki and a business graph, searched in milliseconds, far faster than RAG, with no LLM or database on the read path. HTTP and MCP.