What kind of files does Shinkuro use for prompts?

Shinkuro uses standard markdown files for prompts. Each markdown file represents a single prompt and can optionally include YAML front matter for metadata like 'name' and 'description'.

Shinkuro is an MCP server designed to manage and serve AI prompts. It loads prompts from markdown files stored in local folders or remote Git repositories, streamlining prompt synchronization for developers.

How does Shinkuro load prompts?

Shinkuro loads prompts by reading markdown files. You can specify a local folder path or provide a Git repository URL (supporting GitHub, GitLab, SSH, HTTPS), from which it clones and reads the prompt files.

Does Shinkuro support Git repositories and caching?

Yes, Shinkuro fully supports loading prompts from Git repositories. It caches cloned repositories locally for efficient access and offers an optional `AUTO_PULL` feature to fetch the latest changes automatically.

How do I configure Shinkuro?

Shinkuro is easily configured using environment variables. Key variables include `FOLDER` for local paths, `GIT_URL` for repository links, `CACHE_DIR` to specify cache location, and `AUTO_PULL` for automatic updates.

Shinkuro

Name: Shinkuro
Author: DiscreteTom

•

Manages and serves prompts by loading markdown files from local folders or Git repositories.

Shinkuro

•

Manages and serves prompts by loading markdown files from local folders or Git repositories.

Shinkuro acts as an MCP (Model Context Protocol) server, providing a robust solution for prompt synchronization and management. It intelligently loads markdown files, which are then treated as structured prompts, from either a specified local directory on your filesystem or a remote Git repository. This functionality enables users to centralize, version control, and effortlessly serve their prompt content, ensuring consistency and ease of access for MCP clients within their prompt engineering workflows.

Key Features

01Loads prompts from local markdown files.

02Supports loading prompts from Git repositories (GitHub, GitLab, SSH, HTTPS).

03Caches cloned Git repositories for efficient access.

04Allows optional automatic pulling of latest changes from Git repositories.

05Configurable via environment variables for flexible deployment.

060 GitHub stars

Use Cases

01Centralize and serve prompt content to multiple MCP clients.

02Version control prompt libraries using Git for collaborative development.

03Automate prompt delivery from a shared repository to various applications.