OpenAI Adds Plugin Support to Codex Programming Assistant

OpenAI has introduced a new plugin feature for its Codex programming assistant, enabling developers to customise and extend the tool’s capabilities for specific workflows.

The update allows users to build plugins that combine task-specific “skills” with integrations to external applications. Skills consist of natural language instructions paired with scripts or other technical assets, allowing Codex to execute predefined actions without generating code from scratch each time—reducing errors and improving efficiency.

Plugins can also connect Codex to external services using MCP servers, enabling developers to integrate tools such as development environments or middleware configurations. OpenAI has launched a plugin directory with more than a dozen prebuilt integrations, including support for platforms like Google Drive and GitHub.

The move follows similar functionality introduced by rival Anthropic for its Claude Code platform, highlighting growing competition in AI-powered developer tools. Plugins are expected to simplify collaboration by ensuring consistent configurations across teams, reducing discrepancies in shared codebases.

OpenAI said future updates may expand plugin capabilities further, potentially adding new components beyond current integrations and workflows. The feature could also extend beyond software development, particularly as the company explores broader applications for its AI tools.

The development comes amid reports that OpenAI is working to integrate Codex with ChatGPT into a unified “super app,” which could bring plugin-based customisation to a wider range of use cases, including research and enterprise productivity.