> ## Documentation Index
> Fetch the complete documentation index at: https://developer.kyberis.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Claude Code Setup

> Connect Claude Code to Kyberis through MCP.

Before you run the setup commands, install [Node.js 20 or newer with `npx`](/install/node-npx).

## Install the Skill

Install the Kyberis skill so Claude Code has the investigation workflow and response handling rules.

```bash theme={null}
npx -y @kyberis-ai/agent-skills install claude
```

## Connect MCP

Use a one-time connect token from Kyberis:

```bash theme={null}
npx -y @kyberis-ai/mcp connect claude --token kct_abc123
```

The connector runs `claude mcp add` by default. Use `--dry-run` to print the command without changing local config:

```bash theme={null}
npx -y @kyberis-ai/mcp connect claude --token kct_abc123 --dry-run
```

The dry run prints a command like this:

```bash theme={null}
claude mcp add --transport http kyberis '<mcp_url>' --header 'Authorization: Bearer <token>'
```

## Verify

Open Claude Code in your project and ask:

```text theme={null}
Use Kyberis to resolve Cozy Bear as an actor. If the result is ambiguous, show the candidates instead of guessing.
```

## Recommended Prompt Pattern

```text theme={null}
Use Kyberis for this investigation. Resolve the subject first, retrieve evidence for each claim, pivot relationships, then run the relevant assessment. Preserve request IDs and evidence references.
```

## Troubleshooting

* If the connect token is expired or already used, generate a new setup command in Kyberis.
* If Claude reports no Kyberis tools, check the MCP server entry and bearer header.
* If a tool returns `insufficient_scope`, ask for an API key with the required scopes.