mcp

GibsonAI/mcp

3.5

If you are the rightful owner of mcp and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.

GibsonAI's Model Context Protocol (MCP) server enhances the functionality of various development tools by allowing them to interact with projects, databases, and APIs directly from an IDE.

Tools

Functions exposed to the LLM to take actions

get_projects

Get all of the user's existing GibsonAI projects. If the user mentions a project by name and you don't have the project UUID, call this tool. If there's a .gibsonai file in the project or workspace root directory, assume that the user wants to use that project and don't call this tool unless there's a good reason to do so.

create_project

Create a new GibsonAI project. Before assuming that the user wants to create a new project, look for a .gibsonai file in the project or workspace root directory. If the .gibsonai file exists and contains a project UUID, call get_project_details to get the project details and ask the user if they want to use the existing project. If the .gibsonai file doesn't exist, but the user mentions a project name, call get_projects to check if a project with a similar name already exists and ask the user if they want to use the existing project. If the user explicitly states they want to create a new project, call this tool without confirming. If you call this tool, ask the user if they want to update the .gibsonai file (or create it if it doesn't exist) with the new project UUID.

get_project_details

Get a GibsonAI project's details. If there's a .gibsonai file in the project or workspace root directory, assume that the user wants to use that project unless they explicitly state otherwise.

get_project_hosted_database_details

Get the details for querying a GibsonAI project's hosted database. This includes necessary context for an LLM to understand and generate code related to fetching or modifying the data in the project's hosted database.

update_project

Update a GibsonAI project's details. This currently only updates the project's name. Returns the updated project details.

submit_data_modeling_request

Submit a data modeling request for a GibsonAI project. This tool fully handles all data modeling tasks. The LLM should not attempt to handle any data modeling tasks. If the user describes a data modeling task, call this tool with the user's request as-is. Again, the LLM should never attempt to directly handle data modeling tasks when working on a GibsonAI project. Returns the response from GibsonAI's data modeler.

deploy_project

Deploy a GibsonAI project's hosted databases. This updates both the development and production database schemas simultaneously by automatically handling necessary schema migrations.

get_project_schema

Get the current schema for a GibsonAI project. This includes any changes made to the schema since the last deployment.

get_deployed_schema

Get the deployed schema for a GibsonAI project. This is the database schema that is currently live on the project's hosted databases.

query_database

Query a GibsonAI project's hosted database using SQL. The environment-specific API key must be provided. If you're not sure which environment to use, ask the user for clarification. Always use the correct syntax for the database dialect (found in the project details). Always wrap identifiers in the dialect appropriate quotes (backticks for MySQL, double quotes for PostgreSQL).

Prompts

Interactive templates invoked by user choice

No prompts

Resources

Contextual data attached and managed by the client

No resources