Packages

MCP server for Brex — the corporate card and spend management platform designed for startups and tech companies. Enables AI agents to query Brex transactions, budgets, expenses, and spend analytics. Supports finance automation workflows for startups using Brex as their primary spend management solution.

The standard JSON library for the Typelevel Scala ecosystem. Circe provides type-safe JSON encoding (Encoder[A]) and decoding (Decoder[A]) with automatic derivation via circe-generic. JSON values are represented as algebraic types (Json, JsonObject, JsonArray). Decoders return Either[DecodingFailure, A] — errors are values, not exceptions. Integrates with http4s, Tapir, and most Typelevel libraries as the default JSON codec.

Static code analysis tool for Kotlin — the standard linter and code smell detector for Kotlin projects. Detekt analyzes Kotlin source code for: complexity metrics (cyclomatic complexity, cognitive complexity), code smells (long methods, large classes, magic numbers), potential bugs (null safety violations, empty catch blocks), naming conventions, performance issues, and coroutine misuse. Configurable via detekt.yml, integrates with Gradle/Maven as build task, IDE plugins, and GitHub Actions. Supports custom rule sets for project-specific standards. Replaces Checkstyle for Kotlin codebases.

Facebook Ads Library MCP server enabling AI agents to search and analyze the Facebook Ads Library — querying active and inactive ad campaigns across Meta platforms (Facebook, Instagram), retrieving ad creative details, analyzing competitor advertising strategies, and integrating Meta's public ad transparency data into agent-driven competitive intelligence and market research workflows.

MCP Endpoint Server enabling AI agents to make HTTP requests to arbitrary endpoints — acting as a flexible HTTP client tool that allows agents to call any REST API, webhook, or HTTP service. Provides a general-purpose HTTP request capability without needing a dedicated MCP server for each API, enabling agents to interact with any web service.

MCP server built on Netmiko — the multi-vendor network device SSH automation library — enabling AI agents to connect to and manage network devices including Cisco IOS/NX-OS, Juniper, Arista, Fortinet, and 80+ other network operating systems. Enables agents to send commands, configure interfaces, retrieve device information, and automate network management workflows.

Email verification service that validates individual emails in real-time or processes large lists in bulk, providing bounce prediction and deliverability scores for email marketing hygiene.

MCP server for PTT (批踢踢實業坊) — Taiwan's largest BBS (Bulletin Board System) with over 1.5 million registered users and a central role in Taiwanese internet culture. From the PyPtt organization (official Python PTT client maintainers). Enables AI agents to read PTT boards, search posts, retrieve article content, and monitor Taiwanese online discourse.

Fast, feature-rich Python static type checker from Microsoft. Powers VS Code's Pylance extension. Written in TypeScript for speed — significantly faster than mypy for large codebases. Supports all Python typing features including generics, protocols, TypeVarTuple, ParamSpec, and type narrowing. Can be run standalone or as a language server.

SystemPrompt MCP server enabling AI agents to manage and work with system prompts — creating, storing, and retrieving system prompts, validating prompt safety, managing prompt templates, and integrating prompt management capabilities into agent-driven AI application development and prompt engineering workflows.

Ultimate MCP Server — a comprehensive, batteries-included MCP server providing a wide collection of utilities for AI agents including text processing, file operations, LLM API access, web search, code execution, database queries, and dozens of other tools in a single configurable server, designed as an all-in-one toolkit for agent workflows.

VisionCraft MCP server enabling AI agents to generate images using multiple AI image generation models — integrating Stable Diffusion, FLUX, and other text-to-image models via VisionCraft API, managing generation parameters (resolution, steps, CFG scale, LoRA), and embedding image generation capabilities into agent-driven creative and content workflows.

JavaScript/TypeScript schema validation library with chainable API for defining and validating object shapes. Known for async validation support, excellent Formik integration, and human-readable error messages. Supports nested object schemas, conditional validation, custom transforms, and TypeScript type inference. Widely used in React form validation alongside Formik and React Hook Form.

TypeScript library implementing the Result type pattern from Rust/ML languages. Instead of throwing exceptions, functions return Result<T, E> which is either Ok(value) or Err(error) — forcing callers to explicitly handle both success and failure. Provides a lightweight (< 5KB) alternative to Effect for typed error handling. Includes ResultAsync for Promise-based code. Eliminates try/catch and makes error types visible in function signatures.

Elastic Security REST API and Elasticsearch API for open platform SIEM, XDR, and threat detection. Enables AI agents to manage detection rule creation and management automation, handle alert triage and case management workflows, access threat hunting via Elasticsearch Query Language (EQL), retrieve MITRE ATT&CK aligned detection rule libraries, manage timeline investigation and case collaboration, handle endpoint response actions via Elastic Agent/Fleet, access ML-based anomaly detection rule management, retrieve network and host telemetry from Elastic Agent, manage Kibana SIEM space and access control configuration, and integrate Elastic Security events with SOAR, ticketing, and response platforms.

MCP server for Google Gemini AI image generation capabilities. Enables AI agents to generate images using Google's Gemini/Imagen models via the Gemini API — producing images from text prompts, supporting creative workflows, content generation, and multimodal AI applications.

Debugging print statement replacement for Python — ic() prints variable names, types, and values with context (file, line, parent function). IceCream features: ic(var) prints 'ic| filename.py:10 in func()- var: value', ic() without args prints caller location, ic.disable() to suppress in production, ic.enable() to re-enable, ic.configureOutput() for custom prefix/output function, ic.includeContext for toggling context display, supports any expression not just variables, and returns its argument for inline use. Drop-in upgrade over print() for debugging Python agents.

A cross-platform desktop application that provides a unified dashboard for managing multiple MCP servers — toggling them on/off, grouping them into workspaces, enabling/disabling individual tools, and logging interactions — while keeping all data local.

Flexible Python test automation tool similar to tox but using Python code instead of INI configuration. Nox sessions are Python functions decorated with @nox.session that create virtualenvs, install dependencies, and run test commands. Simpler than tox for complex workflows — run lint, tests, docs, and deployment scripts with one tool. Google uses Nox for many open-source Python projects.

MCP server for Retell AI — a platform for building AI-powered voice agents and phone call automation. Enables AI agents to interact with Retell AI's API — creating voice agents, initiating calls, monitoring call status, retrieving transcriptions, and managing voice AI workflows for customer service, sales, and communication automation.