> For the complete documentation index, see [llms.txt](https://docs.rainbird.ai/rainbird/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.rainbird.ai/rainbird/rainbird-labs/agentic-knowledge-engineering-claude-skill.md).

# Agentic Knowledge Engineering (Claude Skill)

Turn Claude into a Rainbird Knowledge Engineer. Provide policies, rulebooks, or regulations. It researches, designs, builds, tests, and documents production-like Rainbird knowledge graphs.

Today, Co-author creates strong first drafts. It also helps Knowledge Engineers refine and test graphs in Rainbird Studio. Our goal is different: build, test, and deploy production graphs without Rainbird training or manual graph changes.

This research project demonstrates that path.

## What it does

* **Deep research:** Researches the domain, including web searches where needed.
* **Planning:** Proposes a graph architecture and asks clarifying scope questions.
* **Build and validation:** Builds and uploads a knowledge graph. It fixes syntax errors automatically.
* **Live testing:** Creates and runs scenarios through the Rainbird Reasoning Engines API. It refines behaviour where results differ, including adversarial edge-case tests.
* **Documentation:** Detailed documentation that maps requirements to rules.
* **Importable test scenarios:** Returns a JSON test suite for direct import into Rainbird Studio.

## Benefits

**Production-grade knowledge depth:** Deliver deterministic decisions for complex domains with less knowledge-engineering effort and cost.

**Requirement-to-rule traceability:** Capture and number requirements before modelling. References them in rule names and documentation to trace logic back to its source for better auditing.

**Tested live, then refined:** Run scenarios against the live engine. Feed failures back into the model and rerun tests until they pass. Audit explanations as well as outcomes.

**Keep the test suite:** Import the validated scenarios into Studio. Expected values provide regression coverage from day one. Stakeholders can review intended behaviour directly.

## Prerequisites

* A Claude account
* A Rainbird Community Edition account — [sign up](https://app.rainbird.ai/signup)
* The Rainbird Skill ZIP file from the [Rainbird Labs Forum](https://forum.rainbird.ai/c/rainbird-labs/turning-claude-into-an-agentic-knowledge-engineer)

## How to use

1. Download the `rainbird-graph-creation` Skill ZIP file from the [Rainbird Forum](https://forum.rainbird.ai/c/rainbird-labs/turning-claude-into-an-agentic-knowledge-engineer).
2. In Claude, open **Settings → Skills → Add skills**. Upload the ZIP file.
3. Start a new chat. Use a prompt such as: “Using the Rainbird Skill, build a graph that…”
   * Use Opus or Fable for best results. Sonnet is less effective for this complex task.
4. Provide your Community API key when requested.

{% hint style="danger" %}
Only provide your API key when your privacy settings prevent Anthropic using chats for model training.

In Claude, go to **Account → Settings → Privacy**. Ensure **Help improve our AI models** is **off**.
{% endhint %}

## Important considerations

**Create only:** Every graph update creates a new graph in your Rainbird account. This includes syntax fixes, failed-test fixes, and later requests.

**Build time:** A complex build can take one to two hours.

**Data security:** This uses Anthropic’s Claude product and its terms. Do not share personal or sensitive business data without authorisation.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.rainbird.ai/rainbird/rainbird-labs/agentic-knowledge-engineering-claude-skill.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.