SkySpark

You can connect AI assistants to your SkySpark data using the SkySpark MCP Server. It exposes a set of tools to evaluate Axon expressions, read and write SkySpark records, manage functions, and explore projects, all through a consistent MCP interface that works across diverse network setups.

How to use

You interact with SkySpark data by running an MCP client against the SkySpark MCP Server. The server offers multiple transports and supports multi-project access, allowing you to evaluate expressions, read records, manage functions, and explore your SkySpark projects from a single, centralized endpoint.

How to install

Prerequisites: You need Python to install and run the MCP server. You may also need a suitable MCP client to connect to the server.

# Install the SkySpark MCP server package
pip install skyspark-mcp

# Create a configuration file (example values shown; adjust to your environment)
# Create a .env file with SkySpark connection details
# SKYSPARK_HOST=192.168.1.100
# SKYSPARK_PORT=8080
# SKYSPARK_PROJECT=demo
# SKYSPARK_AUTH=scram
# SKYSPARK_USERNAME=su
# SKYSPARK_PASSWORD=your-password

# Run the MCP server with default SSE transport (port 8765)
skyspark-mcp

# Or run with stdio transport for direct MCP client integration
skyspark-mcp --stdio

Authentication

You can authenticate in several ways depending on your network setup and security requirements. The server supports SCRAM by default, no-auth for trusted environments, and token-based authentication.

Multi-Project usage

You can target different SkySpark projects from the same MCP server. Each tool accepts an optional project parameter, so you can run actions against multiple projects from a single endpoint.

Usage with Claude Desktop

Configure Claude Desktop to connect to the SkySpark MCP Server by adding an MCP server entry with the standard command and environment variables. This lets Claude issue MCP requests to SkySpark directly from your Claude workspace.

Usage with mcporter

You can call SkySpark MCP endpoints from mcporter using either the SSE transport or stdio transport. Choose the transport that fits your integration scenario and provide the appropriate command or URL.

Notes on usage

The MCP server is production-tested and designed to work with various network arrangements, including tunnels and VPNs. Base64-safe commits help avoid Axon interpolation issues when committing function changes.

Security and data handling

Maintain secure credentials in your environment. Prefer SCRAM or token-based authentication in exposed networks. For fully trusted environments, no-auth can be used, but you should ensure your network boundary is secure.

Example workflow

1. Install the MCP server and dependencies. 2) Create a .env file with SkySpark connection details. 3) Start the MCP server using the desired transport. 4) Connect your MCP client or orchestration tool to the server and begin evaluating expressions, reading records, and managing functions.

What you can do with the tools

This server exposes a comprehensive set of 21 tools to interact with SkySpark data. You can evaluate Axon expressions, read and write records, manage functions, and navigate your SkySpark projects. Your workflows can read by filter, read by ID, perform full-text search, create and delete records, and manage function sources with base64-safe commits.

Available tools

ss_eval

Evaluate any Axon expression and return the result.

ss_about

Provide server information such as version, project, and user.

ss_read

Read records using Haystack-style filters.

ss_read_by_id

Read a single record by its ID.

ss_search

Perform a full-text search across records.

ss_create

Create a new record in SkySpark.

ss_update

Update record tags.

ss_delete

Delete records matching a filter.

ss_func_list

List project functions.

ss_func_read

Read function source code.

ss_func_commit

Commit updated function source with base64-safety.

ss_func_commit_batch

Batch commit multiple functions.

ss_func_diff

Compare current function source with proposed changes.

ss_funcs_all

List all functions including built-ins.

ss_view_update

Update view source.

ss_page_update

Update Docunomicon page content.

ss_rec_create

Create a record with complex tag values.

ss_views

List views.

ss_templates

List Docunomicon templates.

ss_libs

List installed extensions.

ss_projects

List available SkySpark projects.