Six open-source pieces, one JavaScript agent stack

Most agent projects do not fail because the first model call is difficult. They become difficult when the prototype needs memory, tools, evaluation, a user interface, documentation that coding agents can navigate, and a repeatable review process. That is the problem the open-source AgentsKit ecosystem is trying to solve for JavaScript teams. It is a set of independent projects with shared contracts, rather than one application that must own the whole stack. The six public pieces 1. AgentsKit: the composable foundation AgentsKit provides focused JavaScript and TypeScript packages for runtimes, adapters, tools, skills, memory, RAG, evaluation, observability, sandboxing, and UI bindings. Its core has zero runtime dependencies and a CI-enforced 10 KB gzipped budget. Six formal contracts keep adapters, tools, memory, retrievers, skills, and runtimes substitutable. You can run a first local agent without an API key: npm install @agentskit/core @agentskit/runtime tsx Enter fullscreen mode Exit fullscreen mode Then connect OpenAI, Anthropic, Gemini, Ollama, or another adapter without changing the rest of the runtime composition. 2. AgentsKit Chat: one interaction model, seven renderers AgentsKit Chat defines an interactive agent experience once and renders it through React, React Native, Ink, Vue, Svelte, Solid, or Angular. It also supports deterministic local answers before an optional backend call. That matters for documentation and support interfaces where exact project facts should not require an LLM. pnpm dlx @agentskit/chat-cli@0.4.0 init my-chat --renderer react --yes Enter fullscreen mode Exit fullscreen mode 3. Registry: reusable source, not another dependency The AgentsKit Registry distributes ready-to-use agents in a shadcn-style model: the CLI copies the source into your project, so you can inspect and modify it. npx agentskit add research npx agentskit add code-review Enter fullscreen mode Exit fullscreen mode The public catalog currently includes research, code-review, and knowledge-promotion agents, and is open to contributions. 4. Agents Playbook: engineering rules that can execute Agents Playbook turns engineering practices for coding agents into inspectable patterns, prompts, templates, and zero-dependency gates. Instead of telling an agent to “write production-quality code,” a team can define package boundaries, required evidence, review rules, and executable checks. The repository currently verifies 87 production patterns, six engineering pillars, six SDLC phases, and 13 gate scripts from its own source. 5. Doc Bridge: tell agents where they may work Doc Bridge converts repository documentation and ownership maps into deterministic handoffs: { "startHere": "docs/for-agents/packages/auth.md", "editRoots": ["packages/auth"], "checks": ["pnpm --filter @demo/auth test"], "humanDoc": "/docs/guides/auth" } Enter fullscreen mode Exit fullscreen mode The same handoff is available through CLI, MCP, and CI. The core path requires no LLM or API key. Try the bundled proof: npx -y @agentskit/doc-bridge demo --text Enter fullscreen mode Exit fullscreen mode 6. Code Review CLI: bring the model you already use AgentsKit Code Review runs locally or as a GitHub Action. It supports existing Codex and Claude CLI logins, hosted providers, gateways, and local Ollama models. Seven review lenses generate candidate findings. Independent verification votes then remove weak findings before the result reaches the developer. npx --yes github:AgentsKit-io/code-review-cli --provider codex-cli Enter fullscreen mode Exit fullscreen mode The output can be Markdown, GitHub review comments, or SARIF. How the pieces fit You do not need to adopt everything: Use AgentsKit when you need composable runtime capabilities. Add Chat when the same agent experience must run across web, native, and terminal interfaces. Copy from the Registry when a ready agent is a better starting point than a blank file. Use the Playbook to establish contribution and quality rules. Add Doc Bridge when agents need deterministic ownership, documentation, and validation routes. Run Code Review when agent-authored changes need a low-noise verification layer. Each repository is independently usable and openly licensed. The shared objective is interoperability: a team should be able to start with one package and add another without replacing its existing architecture. Where contributions help most The ecosystem is looking for more than implementation code. Useful contributions include provider adapters, framework examples, reusable agents, false-positive reductions, documentation adapters, executable gates, fixtures, and clearer onboarding. If one of these problems matches work you are already doing, try the smallest relevant project, report what did not compose cleanly, and consider starring the repositories you want other JavaScript developers to discover. Explore AgentsKit Browse the organization Join the Discord