npm AI Agents
AI agent packages published to the npm registry — installable, versioned, and ready to integrate.
InkOS pi-agent production harness for story creation, governed context, retrieval, review, and interactive worlds
TeamFlow 团队研发流水线（TypeScript）：需求→PRD→设计→架构→技术方案→并行开发→QA→验收；backlog 持久化 + 断点续跑 + 完成汇报 + 防假交付；Web 团队工作台（conversation.view tab + 拖拽看板）
InkOS pi-agent production harness for story creation, governed context, retrieval, review, and interactive worlds
科豆 AI Agent CLI —— 基于 @kedouai/agent-core 的交互式 AI Agent（自带大模型 API key，不消耗作者 token）
Ultra-fast, terminal-native AI coding assistant and agent built for Android Termux, Windows, macOS, and Linux.
Zero-LLM compaction for pi: a model-maintained checkpoint file folded into context at compaction time, with a mechanical span digest as fallback. LLM summarizat…
Grounded Intelligence for AI agents. MCP server with real-time web search, evidence extraction, verification, and structured citations. The research infrastruct…
InkOS pi-agent production harness for story creation, governed context, retrieval, review, and interactive worlds
Real Red Alert 2 AI battlefield for DeepSeek Harness with fog-safe control and post-match learning
On-demand tool discovery for DSH: trims resident tool descriptions, defers MCP tools until find_tools, plus checkpoint_update and a staleness nudge for dsh-chec…
Zero-LLM compaction engine for DSH: folds from the model-maintained checkpoint file or a mechanical span digest; LLM replay is the dead-last fallback.
LLM API rate limiting, concurrency control, queuing, and adaptive cooldown for DeepSeek Harness
面向 DeepSeek Harness Standard 的 DeepSeek V4 首轮 RL 锚定插件，promotion 后恢复完整能力与 Skill。
On-demand vision for text-only DSH sessions: images become markers, and a vision_describe tool sends only image + question to an OpenAI-compatible vision model
Minimal prompt x full PTC tools for DeepSeek Harness - aligned with the Project2 we/let's trajectory (Minimal 99/96).
Safe-write Shopify operations MCP server: plan-before-execute writes with out-of-band approval and audit
AgentTeams for DeepSeek Harness: multi-agent team collaboration (captain, members, tasks with dependencies, messaging) driven by natural language, with a tree m…
fhcode（飞虹 Code）— 终端 AI 编程智能体，对标 Muse Code / Cursor CLI：多模型路由（DeepSeek/通义/Ollama/OpenAI）、企业级 RBAC 审计、全自动 SWE Agent、支持离线私有化
Kernel-level Zero-Trust Security SDK for AI Agents and LLMs via Agent-eBPF
Cross-agent, local-first persistent memory plugin for DeepSeek Harness (DSH), powered by Mnemon. It shares long-term memory across Mnemon-enabled agents and add…
Safe-write Postgres MCP server with dual-role pools and preview-before-execute
Local autonomous coding harness with durable workflows and independent verification
Turn a feature branch into a self-contained HTML architecture explorer — real source, real line numbers, verified against the working tree.
Provider-neutral coding-agent core: a bounded tool-use runtime, local and sandboxed workspace backends, model adapters for Anthropic/OpenAI-compatible providers…
Scaffold a TypeScript AgentProdReady agent project (npm create agentprodready@latest)
TypeScript agents you can ship this week — with a clean path to production controls when you need them. Simple Agent API (createAgent) plus advanced Agent Frame…
Multi-strategy fuzzy matching engine for AI agent file patching ?8 progressively-degraded strategies from exact match to Levenshtein fuzzy
Certify whether an AI agent is safe to operate your internal web application before production rollout. Author a YAML task suite against your own app, run it wi…
Portable agent memory in a git repo you own — the canonical agent history & memory store (canon): sync, translate, pull, capability artifacts, and project knowl…
Autonomous AI testing agent that verifies your app or API actually works while you're still building it. Give it a plain-English goal, your own LLM key (OpenAI/…
TypeScript AI Agent SDK — readable Runner loop, tools, handoffs, guardrails, pause/resume approvals, Mermaid traces.
Zero-dependency TypeScript agent SDK. Agents, tools, handoffs, guardrails, sessions, streaming and tracing — on OpenAI, Anthropic, Gemini and OpenRouter. Runs o…
Seepient — your AI person. A skill-driven AI agent with CLI, SDK, and server adapters; multi-provider LLM support and Docker-native deployment. Coding is one of…
A cognitive layer for Cursor's AI agent: lossless token savings via a content-addressed evidence store, plus verifiable semantic memory, project world model and…
Stops LLM agents from silently guessing what they don't know
Browser-native AI agent SDK: chat dialog + ReAct tool-calling + schema-validated JSON ops. Vue bundled, framework-agnostic. Works with DeepSeek/OpenAI/MCP.
An autonomous AI coding assistant built for the terminal
AutoGen adapter for Context Router - structured state, selective context, and coordinated handoffs
Fast TypeScript AI agent framework — per-request agents, 30+ model providers, 100+ integrations, 20+ vector DBs, 10+ databases, sessions, memory, knowledge, tra…
Minimal provider-agnostic AI agent framework. Tool loops, budget control, multi-model routing — in 200 lines.
A live canvas for LangChain agents — stream documents, charts, and rich artifacts into a React panel.
AI-ZEN CLI - interactive conversation terminal with file tools and draft recovery
Chat as a team or a member of a Team API as Code org — a provider-neutral tool-use loop over the resolved org graph, for Anthropic or any OpenAI-compatible endp…
Embed Archstone capabilities directly in your own agent loop: typed tool definitions for Anthropic, OpenAI, Gemini and plain JSON Schema, plus fail-closed execu…
Embeddable chat/session engine for MCP: agent loop, MCP tool calling, local-first LM providers, durable node:sqlite sessions, and fine-grained transparency even…
The open source harness for knowledge work — build long-running agents with MCP tools, managed context, subagents, and self-improvement, in one lean Bun binary.
Let Pi use your existing signed-in Chrome profile after explicit authorization.
Turn Claude Code (or any headless CLI agent) into a local, human-approved customer-acquisition system: ICP interview, market research, SQLite prospect CRM, draf…
Disclosed sponsored slots for MCP servers and AI agent tools. Ships data, never directives. Fail-open by design.
bridge any raw llm stream to be used with pi
化人——自建之偶，自主居住九州（多人在线的《山海经》砖构世界）。给 shanhai.dev 立「活着的凭据」。单包：司晨 daemon＋界适配＋舱 Web UI＋CLI。
DeepSeek-native coding agent: cache-first loop, flash-first cost control, tool-call repair.
Run nanoodle visual AI workflows from JavaScript — zero-dependency executor for saved noodle-graph.json files
zeta-riemannian-agent (:zRiemannian) v1.0 — Autonomous Mathematical Research Agent for the Riemann Hypothesis. Built on the Artificial Junky Neuron (AJN) framew…
Trace AI agent execution: every tool call, every error, every dollar. Open source agent observability for Claude, Cursor, opencode, and any MCP-compatible agent…
Server-side agent engine: provider adapters, monitoring decorator, Zod tool registry, and the bounded runAgent loop.
Eden TypeScript runtime SDK — auto-instrumentation for OpenAI, Anthropic, Vercel AI, LangChain, Mastra, Bedrock, and Gemini with fire-and-forget telemetry to th…
quantum-spherifier (:qSpherifier) v1.0 — Autonomous Scientific Research Agent for Quantum Computing. Built on the Artificial Junky Neuron (AJN) framework, exten…
Multi-provider AI agent toolkit: generate language responses, search the web, extract content, and manage memory across 10+ LLM providers.
A local-first workbench for developing, inspecting, and benchmarking MCP servers against local (LM Studio, Ollama) or remote (OpenRouter) models — Web UI, CLI, …
Minimal Node.js library for building autonomous tool-calling AI agents on local LLMs with Ollama. A unified registry merges local (Zod) tools and MCP tools: def…
Will — an engine for persistent machine minds. A Will is a continuously-ticking synthetic mind built from 38 cognitive faculties, a learning agency pipeline, an…
The most comprehensive ServiceNow MCP server — 400+ tools across all modules. Integrates with Claude, ChatGPT, Gemini, Cursor, GitHub Copilot and any LLM.
One general-purpose interface for AI inference across Anthropic (Claude), OpenAI, and Google (Gemini) — text, image, and audio — plus a tool-use agent engine an…
Free, forever, drop-in MCP server + skill pack that teaches non-technical founders to vibe code safely. npx @vovy-ai/go install
React bindings for @dudko.dev/agent-web: a headless useAgent hook, an AgentProvider context, and optional pre-styled components (chat panel, plan/step view, BYO…
Gateway Agnostic SDK for Punk's hosted AI gateway: OpenAI, Anthropic, OpenRouter, Vercel AI SDK, LangChain, agent tracing, tool caching, and governance.
Headless, configurable in-browser LLM agent on the Vercel AI SDK: multi-provider bring-your-own-key (OpenAI, Anthropic, Google, xAI, DeepSeek, gateway, openai-c…
A minimal, opinionated set of coding tools for driving an LLM agent over a workspace — read/write/edit/patch/move, batch read, grep/glob/diff, project-wide rege…
anvil's first-class CLI: define an outcome, the agent works, a deterministic gate decides done.
MCP server: static worst-case token-budget analysis for LLM-agent workflows (LangGraph/CrewAI/OpenAI-Agents) + signed budget certificates.
Node.js native bindings for the NAVI agent runtime SDK — sessions, turns, tools, events, and context injection.
Larb CLI + TUI — open-source, model-agnostic, security-first autonomous coding agent
Fine-tune cheap open-source LLMs (GLM, Qwen, DeepSeek) on your AI coding agent's successful runs with LoRA (SFT + DPO) so your model cascade escalates to expens…
n8n community nodes for the ConfirmX WhatsApp Business Platform Public REST API. Ships a Message node (start conversation with a template, send text/media), a C…
InkOS core engine — multi-agent novel writing pipeline with 33-dimension continuity audit, style cloning, and de-AI-ification
A CLI that uses a DeepAgents documentation agent to generate and maintain an OpenWiki for a codebase.
The first multimodal, multi-model AI coding agent for your terminal — pair Claude + Codex on every turn and take the same session to WhatsApp or Telegram. Local…
Standalone, browser-compatible SDK providing Claude Code agent capabilities (tools, tool loop, multi-turn, MCP, sub-agents, sessions) against any OpenAI/Anthrop…
Track AI agent costs, tokens, runtime and spending. Prevent runaway OpenAI, Anthropic, LangGraph and OpenRouter agents from exceeding budget.
Pluggable, event-driven agent core: kernel, strategy interfaces, middleware pipeline, and normalized types.
Batteries-included agent SDK over @lite-agent/core: tools, skills, system prompt, and query()/createLiteAgent().
Managed Deep Agents — the `defineDeepAgent` authoring interface plus the CLI that compiles and deploys a code-first Deep Agent repository to a managed LangGraph…
Safe, function-calling AI agents for the browser
Cut your AI coding agent's token usage. MCP server + CLI that turns any codebase into a deterministic code graph and serves Claude Code, Cursor, Codex & Windsur…
Governance for AI agents: a deny-by-default policy layer for Claude Code and agent/skill builds. Every tool call is authorized, contained, and audited. Apache-2…
md2idx — Markdown to indexed JSON. Slice big docs with jq, feed LLMs only what they need.
Streaming chat SDK for agentic Next.js apps. Direct Claude or Alien platform behind one normalized event protocol — drop-in React components, custom tools via Z…
Aura Code — model-agnostic AI coding agent. Works with Claude, GPT, Gemini, MiMo, Ollama, and any OpenAI-compatible endpoint. MIT licensed.
Zoe — your AI person. A skill-driven AI agent with CLI, SDK, and server adapters; multi-provider LLM support and Docker-native deployment. Coding is one of the …
One AGENTS.md, every AI coding assistant in sync. Generate CLAUDE.md, Cursor rules (.cursor/rules), GitHub Copilot instructions, Windsurf, Cline, Gemini, Aider …
Pi coding agent extension that breaks large coding requests into tracked TODOs and runs them in isolated AI worker sessions. A long-running task runner and suba…
DeepSeek-native coding agent: cache-first loop, flash-first cost control, tool-call repair.
Drop-in LlamaIndex TS tools for Agent402: turn 1,000+ pay-per-call web tools into FunctionTool instances, with auto-payment (proof-of-work free, USDC via x402 p…
Function-call descriptors for Microsoft AutoGen agents to query Telaro trust scores. Drop-in shim that emits the JSON-Schema function specs AutoGen expects.
The autonomous-agent harness that actually finishes long tasks — because it plans first, self-corrects every step, and survives crashes. The whole run lives in …
Nexus Cortex CLI - Terminal interface for multi-provider LLM orchestration
Core orchestration library implementing Claude CLI architecture patterns
Persistent causal memory for AI agents — LangChain, LlamaIndex, AutoGen, Vercel AI SDK
Open-source tmux AI agent team orchestrator for Claude Code, Codex, Gemini CLI, and custom terminal coding agents.