Jilebi Plugins

January 1, 2026 ยท View on GitHub

Jilebi is a Model Context Protocol (MCP) runtime that provides secure sandboxing for MCP plugins, ensuring they can only access resources and capabilities that users explicitly allow. This repository contains a collection of plugins that extend Jilebi's functionality across various domains including development tools, documentation access, file management, and more.

Installation

To install any of these plugins in your Jilebi runtime:

jilebi plugins add <plugin_name>

For example:

jilebi plugins add github
jilebi plugins add filesystem
jilebi plugins add context7

Manifest.toml Format

Each plugin includes a manifest.toml file that defines its capabilities, permissions, and interface. The manifest follows this schema:

Basic Structure

name = "plugin-name"           # Plugin identifier (required)
version = "1.0.0"             # Plugin version
homepage = "https://..."       # Plugin homepage URL
creator = "Plugin Creator"     # Plugin creator/maintainer
contact = "support@..."        # Contact information

[env]                         # Environment variables (optional)
VAR_NAME = { schema = { type = "string" }, default = "value" }

[secrets]                     # Secrets/API keys (optional)
API_KEY = { schema = { type = "string" } }

[resources]                   # Static resources (optional)
[resources.resource-name]
name = "Resource Display Name"
description = "What this resource provides"
mime_type = "text/plain"
function = "function_name"
[resources.resource-name.permissions]
# Permission settings...

[tools]                       # Interactive tools (required)
[tools.tool-name]
name = "tool-name"
description = "What this tool does"
input_schema = { type = "object", properties = { ... }, required = [<keys>] }
function = "function_name"
[tools.tool-name.annotations]
title = "Tool Display Title"
read_only_hint = true/false
destructive_hint = true/false
idempotent_hint = true/false
open_world_hint = true/false
[tools.tool-name.permissions]
# Permission settings...

[prompts]                     # Conversation templates (optional)
[prompts.prompt-name]
name = "prompt-name"
description = "What this prompt helps with"
arguments = [
  { name = "arg1", description = "Argument description", required = true },
]
messages = [
  { role = "user", content = { type = "text", content = "Template content" } },
]

Permission Types

  • hosts - Array of allowed host domains
  • urls - Array of specific allowed URLs
  • read_files - Array of file paths/patterns the plugin can read
  • write_files - Array of file paths/patterns the plugin can write
  • read_dirs - Array of directory paths the plugin can read from
  • write_dirs - Array of directory paths the plugin can write to

Available Plugins

GitHub Plugin

The github plugin was derived from the original MCP server

Tools: Repository management, issue tracking, pull requests, workflow automation
Resources: None
Prompts: issue-to-fix-workflow
Permissions: hosts: ["https://api.github.com"]
Secrets: GITHUB_TOKEN

Provides comprehensive GitHub integration including repository search, issue management, pull request creation, commit history, file content access, workflow monitoring, and gist creation.

Key Tools:

  • get-me - Get authenticated user information
  • search-repositories - Search GitHub repositories
  • create-issue - Create new issues
  • create-pull-request - Create pull requests
  • get-file-contents - Read repository files
  • list-commits - Browse commit history
  • create-gist - Create code gists

Fetch Plugin

Tools: URL fetching and content extraction
Resources: None
Prompts: None
Permissions: hosts: ["https://*", "http://*"]
Secrets: None

Fetch web content from URLs and convert HTML pages to clean, readable markdown format.

Key Tools:

  • fetch_url - Fetch a URL and return content as markdown or raw text

Parameters:

  • url (required) - The URL to fetch (HTTP or HTTPS)
  • max_length (optional) - Maximum characters to return (default: 5000, max: 1000000)
  • start_index (optional) - Starting index for pagination (default: 0)
  • raw (optional) - Return raw HTML instead of markdown (default: false)

Features:

  • Automatic HTML to markdown conversion
  • Body content extraction (removes head, scripts, styles)
  • Content pagination for large pages
  • Raw mode for unprocessed HTML

Filesystem Plugin

The filesystem plugin was derived from the filesystem MCP implementation

Tools: File and directory operations
Resources: None
Prompts: None
Permissions: read_dirs: ["user_defined"], write_dirs: ["user_defined"]
Secrets: None

Comprehensive file system operations for reading, writing, and managing files and directories with sandbox protection.

Key Tools:

  • read-text-file - Read text files with head/tail support
  • read-media-file - Read images/audio as base64
  • write-file - Create/overwrite files
  • edit-file - Make selective edits using pattern matching
  • create-directory - Create directories
  • list-directory - List directory contents with sizes
  • directory-tree - Get recursive directory structure
  • move-file - Move/rename files and directories
  • search-files - Find files matching patterns

Context7 Plugin

original MCP server

Tools: Documentation access and library discovery
Resources: None
Prompts: library-summary
Permissions: hosts: ["https://context7.com"]
Secrets: None

Access up-to-date documentation for popular libraries and frameworks through the Context7 service.

Key Tools:

  • resolve-library-id - Find Context7-compatible library identifiers
  • get-library-docs - Fetch comprehensive library documentation

Cloudflare Plugin

original MCP server

Tools: Cloudflare service management and documentation
Resources: None
Prompts: workers-prompt-full, cloudflare-optimization
Permissions: hosts: ["https://api.cloudflare.com", "https://developers.cloudflare.com"]
Environment: CLOUDFLARE_ACCOUNT_ID
Secrets: CLOUDFLARE_API_TOKEN

Manage Cloudflare services including Workers, AI Gateway, and access comprehensive documentation.

Key Tools:

  • ai-gateway-list-gateways - List AI Gateway instances
  • workers-list-scripts - List Cloudflare Workers scripts
  • browser-get-html - Get HTML content using Cloudflare Browser API
  • docs-search - Search Cloudflare developer documentation

Memory Plugin

original MCP server

Tools: Knowledge graph management
Resources: None
Prompts: None
Permissions: None
Secrets: None

Create and manage a persistent knowledge graph for storing and retrieving interconnected information.

Key Tools:

  • create-entities - Create new entities in the knowledge graph
  • add-observations - Add observations to existing entities
  • create-relations - Create relationships between entities
  • read-graph - Read the entire knowledge graph
  • search-nodes - Search entities and observations
  • open-nodes - Retrieve specific entities by name
  • delete-entities - Remove entities and their relations
  • delete-observations - Remove specific observations
  • delete-relations - Remove relationships

AniList Plugin

original MCP server

Tools: Anime and manga database access
Resources: None
Prompts: anime-recommendation, character-analysis, seasonal-anime
Permissions: hosts: ["https://graphql.anilist.co"]
Secrets: None

Access comprehensive anime and manga information from the AniList database.

Key Tools:

  • search-anime - Search for anime titles
  • search-manga - Search for manga titles
  • search-characters - Search for anime/manga characters
  • get-anime-details - Get detailed anime information
  • get-manga-details - Get detailed manga information
  • get-character-details - Get character information
  • get-trending-anime - Get currently trending anime
  • get-trending-manga - Get currently trending manga

Rust Docs Plugin

original MCP server

Tools: Rust crate documentation and discovery
Resources: None
Prompts: rust-documentation-guide, crate-comparison, rust-learning-path
Permissions: hosts: ["https://crates.io", "https://docs.rs"]
Secrets: None

Search and access Rust crate documentation and examples.

Key Tools:

  • search-rust-docs - Search across multiple Rust crates
  • get-crate-info - Get basic crate information and versions

Met Museum Plugin

original MCP server

Tools: Metropolitan Museum of Art collection access
Resources: None
Prompts: explore-art-collection, art-discovery, museum-educator
Permissions: hosts: ["https://collectionapi.metmuseum.org"]
Secrets: None

Explore and search the Metropolitan Museum of Art's collection.

Key Tools:

  • search-museum-objects - Search the museum collection
  • get-museum-object - Get detailed object information
  • list-departments - List museum departments

Time Plugin

Tools: Time zone conversion and current time
Resources: None
Prompts: None
Permissions: None
Secrets: None

Handle time zone conversions and get current time in different zones.

Key Tools:

  • get-current-time - Get current time in specific timezone
  • convert-time - Convert time between timezones

Sequential Thinking Plugin

original MCP server

Tools: Structured problem-solving framework
Resources: None
Prompts: None
Permissions: None
Secrets: None

Provides a framework for dynamic and reflective problem-solving through sequential thoughts.

Key Tools:

  • sequential-thinking - Structured thinking process with branching and revision capabilities

Jilebi Gen Plugin

Tools: Plugin generation and development assistance
Resources: Manifest schema, examples
Prompts: create-plugin
Permissions: None
Secrets: None

Generate new Jilebi plugins with proper structure and documentation.

Key Tools:

  • Various generation tools for creating plugin boilerplates and code

Grafana Plugin

original MCP server

Tools: Comprehensive Grafana monitoring and observability management
Resources: None
Prompts: None
Permissions: hosts: ["user_defined"] (configurable Grafana instance)
Environment: GRAFANA_URL, GRAFANA_USERNAME, GRAFANA_PASSWORD
Secrets: GRAFANA_SERVICE_ACCOUNT_TOKEN, GRAFANA_API_KEY

Complete Grafana integration for dashboards, data sources, alerting, incident management, and observability with support for Prometheus, Loki, and OnCall.

Key Tools:

  • Dashboard Management: get-dashboard-by-uid, update-dashboard, search-dashboards
  • Data Source Operations: get-datasource-by-uid, get-datasource-by-name
  • Prometheus Integration: query-prometheus, list-prometheus-metric-metadata, list-prometheus-metric-names, list-prometheus-label-names, list-prometheus-label-values
  • Loki Log Analysis: query-loki-logs, query-loki-stats, list-loki-label-names, list-loki-label-values
  • Alerting: list-alert-rules, get-alert-rule-by-uid, list-contact-points
  • Incident Management: list-incidents, create-incident, get-incident, add-activity-to-incident
  • OnCall Management: list-oncall-schedules, get-oncall-shift, get-current-oncall-users
  • System Operations: get-health, get-version, search

Wikipedia Plugin

original MCP server

Tools: Wikipedia article access, search, and content analysis
Resources: None
Prompts: None
Permissions: hosts: ["https://wikipedia.org"]
Environment: WIKIPEDIA_LANGUAGE (default: "en"), WIKIPEDIA_COUNTRY, WIKIPEDIA_ENABLE_CACHE (default: false)
Secrets: WIKIPEDIA_ACCESS_TOKEN
Original Server: wikipedia-mcp

Access and analyze Wikipedia content with support for searching, reading articles, extracting facts, and exploring related topics.

Key Tools:

  • search-wikipedia - Search Wikipedia for articles matching a query
  • get-article - Get the full content of a Wikipedia article
  • get-summary - Get a concise summary of a Wikipedia article
  • summarize-article-for-query - Get a summary tailored to a specific query
  • summarize-article-section - Summarize a specific section of an article
  • extract-key-facts - Extract key facts from articles, optionally focused on a topic
  • get-related-topics - Discover topics related to an article via links and categories
  • get-sections - Get the table of contents/sections of an article
  • get-links - Get all links contained within an article
  • get-coordinates - Get geographic coordinates associated with an article
  • test-wikipedia-connectivity - Diagnostic tool for Wikipedia API connectivity

Nix Plugin

original MCP server

Tools: NixOS packages, options, Home Manager, nix-darwin, flakes, and NixHub version tracking
Resources: None
Prompts: None
Permissions: hosts: ["https://search.nixos.org", "https://nix-community.github.io", "https://nix-darwin.github.io", "https://www.nixhub.io"]
Environment: ELASTICSEARCH_URL (optional, default: "https://search.nixos.org/backend")
Secrets: None

Comprehensive NixOS ecosystem integration providing access to 130K+ packages, 22K+ configuration options, Home Manager settings, nix-darwin macOS configurations, community flakes, and package version history via NixHub.

Key Features:

  • Real-time data from official NixOS Elasticsearch APIs
  • Dynamic channel discovery and resolution (stable always points to current stable release)
  • Smart option suggestions when exact matches aren't found
  • Version-aware package searches with nixpkgs commit hashes
  • Category browsing for organized option exploration
  • Cross-platform compatibility (works on any system)
  • Comprehensive console.log debugging throughout

NixOS Tools:

  • nixos-search - Search packages, options, programs, or flakes with channel selection
  • nixos-info - Get detailed information about packages or options
  • nixos-channels - List available NixOS channels with status and document counts
  • nixos-stats - Get package and option counts for channels

Home Manager Tools:

  • home-manager-search - Search Home Manager configuration options
  • home-manager-info - Get detailed option information with type and description
  • home-manager-stats - Statistics about available options and categories
  • home-manager-list-options - List all option categories with counts
  • home-manager-options-by-prefix - Explore options by prefix (e.g., "programs.git")

nix-darwin Tools:

  • darwin-search - Search macOS configuration options
  • darwin-info - Get detailed nix-darwin option information
  • darwin-stats - Statistics about available darwin options
  • darwin-list-options - List all darwin option categories
  • darwin-options-by-prefix - Explore darwin options by prefix (e.g., "system.defaults")

Flakes Tools:

  • nixos-flakes-stats - Statistics about available community flakes
  • nixos-flakes-search - Search community NixOS flakes by name or description

NixHub Tools:

  • nixhub-package-versions - Get version history with nixpkgs commit hashes (up to 50 versions)
  • nixhub-find-version - Find specific package version with commit hash for pinning

Data Sources:

AWS Plugins

A suite of plugins for interacting with Amazon Web Services (AWS). These plugins were migrated from the AWS MCP Servers and provide comprehensive AWS functionality.

AWS Documentation Plugin (aws-docs)

Tools: AWS documentation access and search
Resources: None
Prompts: documentation-lookup
Permissions: hosts: ["https://docs.aws.amazon.com", "https://proxy.search.docs.aws.amazon.com", "https://contentrecs-api.docs.aws.amazon.com"]
Secrets: None

Access AWS documentation, search for content, and get recommendations through official AWS documentation APIs.

Key Tools:

  • read-documentation - Fetch and convert AWS documentation pages to markdown
  • search-documentation - Search AWS documentation using the official search API
  • get-recommendations - Get related content recommendations for documentation pages

Security Model

Jilebi's sandbox architecture ensures that:

  1. Resource Access Control: Plugins can only access files, directories, and network resources explicitly permitted in their manifest
  2. Secret Management: API keys and sensitive data are managed securely and only provided to authorized plugins
  3. Permission Isolation: Each plugin operates in its own security context
  4. User Consent: Users must explicitly approve plugin permissions before installation

Contributing

To contribute a new plugin:

  1. run the command jilebi plugin create
  2. Follow the steps and create a new plugin
  3. Define the plugin interface in manifest.toml
  4. Implement your plugin logic in index.js or index.ts
  5. Include a README.md with usage instructions
  6. Add appropriate package.json and build configuration

The workflow will automatically package and distribute your plugin when merged.