OpenAI Launches 20 Codex Plugins Connecting Slack, Figma, and Productivity Tools

OpenAI has officially launched the Codex Plugin Directory, featuring 20 integrations designed to transform the coding assistant into an autonomous workflow engine. This update allows developers to connect their environments directly to Slack, Figma, GitHub, and Google Drive, facilitating a more integrated software development lifecycle. The rollout represents a strategic move toward multi-agent orchestration within the engineering workspace.

The launch accompanies a new standalone Codex app for macOS and Windows that supports sophisticated background operations. By utilizing Git worktrees, the platform can now handle parallel tasks, such as refactoring authentication while simultaneously building new features in separate environments. This functionality prevents state collisions and significantly accelerates team velocity.

Plugins bundle three distinct components: skills, app integrations, and Model Context Protocol (MCP) servers. Skills provide context-specific instructions for complex work, while app integrations give Codex the ability to read and write data within external productivity tools. MCP servers act as bridges to external databases and systems outside the local project boundary.

A significant addition to the ecosystem is the Automations feature, which allows Codex to perform background tasks without direct user prompts. This includes routine engineering chores such as triaging GitHub issues, monitoring CI/CD pipelines, and resolving system alerts. Developers can now delegate low-level maintenance to focus on higher-level architectural decisions.

Developers can access these new tools through a curated directory in the Codex app or via the command line interface. Once installed, users can invoke specific tools using the "@" symbol or allow the AI to choose the best plugin based on natural language requests like "Summarize unread Slack threads." Detailed documentation on these workflows is available at the OpenAI Codex guide. To maintain enterprise standards, OpenAI has implemented strict governance controls for these new integrations. Administrators can manage plugin access through role-based permissions, ensuring that data sharing remains compliant with internal security policies. All connected third-party services continue to be governed by their respective privacy and authentication protocols.

OpenAI has officially launched its Codex Plugin architecture, debuting a curated directory of 20 official plugins to enhance its agentic capabilities. This major update allows the 1.6 million weekly active users to connect their AI assistant directly to essential workplace tools like Slack, Figma, and Notion. By bundling skills, app integrations, and Model Context Protocol servers, Codex now functions as a comprehensive assistant capable of complex team coordination.

Access and Integration

Users can access these new features through the standalone Codex app, the command line interface, or popular IDE extensions. Installation is streamlined via a central directory where developers can browse curated options or use a simple CLI command. Once installed, the assistant can implicitly select the right tool based on a natural language prompt or be explicitly triggered using specific commands for tighter control.

Enhanced Skills and Security

Beyond simple connections, plugins include "Skills" which are reusable instructions that guide Codex through specialized workflows. Security remains a priority as all integrations respect existing user approval settings and third-party authentication protocols. This ensures that data shared with external apps remains subject to their respective privacy policies and terms of service.

Empowering Custom Workflows

For those building custom tools, OpenAI provided new scaffolding skills to help developers create their own manifests and configurations. This allows engineering teams to standardize local environments by packaging internal documentation and APIs into shareable bundles. By moving away from manual configuration files, the platform makes advanced data connections discoverable and reproducible for entire organizations. Engineering teams can explore the full range of integrations and documentation at the Codex developer portal. The release marks a significant milestone in making AI agents more useful in professional environments.