# Sikaru SDK

> Sikaru SDK documentation for managed AI agents that continually learn from production failures, user feedback, evals, Harbor-compatible agent trajectories, and OpenInference-compatible traces.

## Docs

- [Sikaru SDK: Continual Learning for Managed Agents](https://www.sikaru.ai/docs/index.md): Build managed AI agents that continually improve from production failures, user feedback, evals, Harbor-compatible trajectories, and OpenInference-compatible traces.
- [Core Concepts for Managed Agents](https://www.sikaru.ai/docs/concepts.md): How Sikaru uses continual learning, managed AI agent runs, evals, Harbor trajectories, and OpenInference-compatible traces.
- [Quickstart](https://www.sikaru.ai/docs/quickstart.md): Install the SDK, configure a project key, declare a tool, and run your first Sikaru-managed agent with Experience capture for continual learning.
- [Managed Agents, Runs, and Tools](https://www.sikaru.ai/docs/agents.md): Run Sikaru-managed agents, stream durable runs, resume executions, and register product-owned tools for continual learning workflows.
- [Experience, Harbor Trajectories, and OpenInference](https://www.sikaru.ai/docs/experience.md): Record Sikaru Experience trajectories, upload Harbor ATIF agent trajectories, and ingest OpenInference-compatible trace data for continual learning and eval evidence.
- [Standards-Native Agents and Evals](https://www.sikaru.ai/docs/standards.md): Define managed agent behavior with AGENTS.md, Markdown skills, eval rubrics, production traces, and explicit improvement objectives.
- [Review Continual Learning Improvements](https://www.sikaru.ai/docs/improvements.md): Inspect evidence-backed behavior updates from failures, feedback, traces, and evals, then approve managed-agent improvements to staging and production.

## Optional

- [Sikaru](https://sikaru.ai)
- [Platform](https://sikaru.ai/platform)