Can I use Anthropic as a CLI?

Yes. KosmoKrator exposes a Anthropic CLI through kosmo integrations:call and the provider shortcut command. Commands can return JSON for scripts, CI, and coding agents.

Can I expose Anthropic through MCP?

Yes. The KosmoKrator MCP gateway can expose selected Anthropic tools to clients such as Claude Code, Cursor, Codex, and other MCP-compatible tools.

Where are Anthropic credentials stored?

Anthropic credentials are configured in KosmoKrator's local integration settings and reused by CLI, Lua, and MCP gateway surfaces.

Can I deny write access for Anthropic MCP tools?

Yes. Start the gateway with --write=deny for read-only clients, or use ask/allow only for trusted workspaces.

Anthropic MCP Integration for Claude Agent SDK

Connect Anthropic to Claude Agent SDK through the local KosmoKrator MCP gateway with scoped tools, credentials, and write policy.

7 functions 6 read 1 write API key auth

Submit Anthropic feedback View package source

Connect Anthropic to Claude Agent SDK

Give Claude Agent SDK workflows access to KosmoKrator integrations through a local MCP server.

Add a KosmoKrator stdio MCP server to the Claude Agent SDK options. The gateway is local, scoped to this integration, and starts with --write=deny so Claude Agent SDK can inspect read-capable tools without receiving write access by default.

Anthropic MCP Config for Claude Agent SDK

Use a narrow integration list so the agent does not load unrelated tools.

{
  "mcpServers": {
    "kosmokrator-anthropic": {
      "type": "stdio",
      "command": "kosmo",
      "args": [
        "mcp:serve",
        "--integration=anthropic",
        "--write=deny"
      ]
    }
  }
}

Run the Gateway Manually

kosmokrator mcp:serve --integration=anthropic --write=deny

Why Use KosmoKrator Here

Scoped tools

Expose only Anthropic instead of a broad multi-service tool list.

Local credentials

Reuse credentials already configured for the KosmoKrator CLI and Lua runtime.

Write policy

Start read-only, then opt into ask or allow for trusted workspaces.

Anthropic Tools Visible to Claude Agent SDK

Claude Agent SDK sees stable MCP tool names generated from the Anthropic integration catalog.

MCP tool	Source function	Type	Description
`integration__anthropic__anthropic_list_messages`	`anthropic.anthropic_list_messages`	Read	List messages in the Anthropic conversation history. Returns paginated message resources with optional filtering by model, date, and ID.
`integration__anthropic__anthropic_create_message`	`anthropic.anthropic_create_message`	Write	Send a prompt to Claude and receive an AI-generated response. Supports multi-turn conversations, system prompts, temperature control, and configurable output length.
`integration__anthropic__anthropic_list_models`	`anthropic.anthropic_list_models`	Read	List available Anthropic AI models. Returns model identifiers, creation dates, and display names.
`integration__anthropic__anthropic_get_model`	`anthropic.anthropic_get_model`	Read	Get detailed information about a specific Anthropic model, including its display name, creation date, and capabilities.
`integration__anthropic__anthropic_list_workspaces`	`anthropic.anthropic_list_workspaces`	Read	List Anthropic workspaces. Returns workspace identifiers, names, and configuration details.
`integration__anthropic__anthropic_get_workspace`	`anthropic.anthropic_get_workspace`	Read	Get detailed information about a specific Anthropic workspace, including its name, configuration, and usage settings.
`integration__anthropic__anthropic_get_current_user`	`anthropic.anthropic_get_current_user`	Read	Get the authenticated user's profile and account information.

Related Anthropic Pages

Anthropic CLIUse direct JSON commands instead of MCP discovery. Anthropic MCPFull gateway reference and MCP tool names. KosmoKrator for Claude Agent SDKClient-level gateway setup.