How to Get Support for OpenAI Plugin Development Issues

Contact OpenAI's support team via the email field in your plugin's manifest, leverage the built-in plugin-creator skill, or open a GitHub issue in the openai/plugins repository.

When building extensions for the OpenAI ecosystem, knowing how to access official plugin development support ensures you can resolve blocking issues quickly. The openai/plugins repository provides multiple structured channels for assistance, from direct email contacts embedded in every plugin manifest to automated scaffolding tools. Below are the four authoritative methods to get help, backed by the repository's actual source code and configuration files.

Contact OpenAI Support via the Plugin Manifest

Every plugin in the repository includes a mandatory email field in its manifest file that points to the official support address. This field is located at plugins/{plugin-name}/.codex-plugin/plugin.json and serves as the primary escalation path for developers.

For example, the Zotero plugin's manifest contains:

{
  "name": "zotero",
  "email": "[email protected]"
}

This pattern is consistent across all plugins, including the Teams plugin and Slack plugin. When you encounter authentication errors, API compatibility issues, or manifest validation failures, sending a detailed description to [email protected] routes your request directly to the OpenAI plugin engineering team.

Use the Plugin-Creator Skill for Development Guidance

The repository ships with an automated agent called plugin-creator designed to walk you through common development tasks. This skill helps scaffold new plugins, validate manifests, and troubleshoot local testing errors.

The skill's configuration is defined in .agents/skills/plugin-creator/agents/openai.yaml, which specifies the OpenAI model parameters and includes a fallback contact entry pointing to the support email. To invoke the helper script for generating a properly configured plugin structure, run:

python .agents/skills/plugin-creator/scripts/create_basic_plugin.py \
    --name my-awesome-plugin \
    --description "My custom plugin" \
    --api-type openapi \
    --api-url https://my-awesome-plugin.com/openapi.json

This script automatically injects "email": "[email protected]" into the generated plugin.json, ensuring your support contact is never omitted during initial setup.

Open a GitHub Issue for Bugs and Feature Requests

For reproducible bugs, unexpected API behavior, or missing documentation, filing a public issue in the openai/plugins repository allows the engineering team to track and prioritize fixes. The team monitors issues at https://github.com/openai/plugins/issues and will triage reports that include specific plugin names and error logs.

When submitting an issue, reference the specific file path where the error occurred. For instance, if you encounter an unsupportedApis error while using the Zoom Video SDK on Windows, include the example file path:

**Title:** Zoom Video SDK – `unsupportedApis` error on Windows

**Description:**  
When initializing the Zoom Video SDK on Windows 10, the client returns:

unsupportedApis: ["runRenderingContext"]

Steps to reproduce:
1. Run `plugins/zoom/skills/video-sdk/windows/examples/send-raw-video.md` on a fresh Windows install.
2. Observe the console output.

**Expected behavior:**  
The SDK should initialize without `unsupportedApis` on a supported client version.

**Environment:**  
- Windows 10 Pro, build 19044  
- Zoom Video SDK v2.4.1  

*Contact:* [email protected]

Include your plugin manifest and relevant stack traces to expedite resolution. The OpenAI team may reply directly in the thread or contact you via the email address listed in your issue.

Reference the Repository README

The top-level README.md file provides an overview of the repository structure and explicitly documents the support email convention. Before escalating an issue, consult this file to verify your plugin structure matches the expected format, as many support requests stem from malformed manifest files or incorrect directory layouts.

Summary

• Email Support: Every plugin.json manifest includes "email": "[email protected]" as the canonical contact for direct assistance.
• Automation: The plugin-creator skill in .agents/skills/plugin-creator/agents/openai.yaml scaffolds plugins with support contacts pre-configured.
• Public Issues: File detailed bug reports at github.com/openai/plugins/issues including file paths like plugins/zoom/skills/video-sdk/windows/examples/send-raw-video.md.
• Documentation: The README.md explains repository layout and support conventions for self-service troubleshooting.

Frequently Asked Questions

What is the official support email for OpenAI plugin development?

The official support email is [email protected]. This address is hardcoded into every plugin manifest file (e.g., plugins/zotero/.codex-plugin/plugin.json) and serves as the primary channel for technical assistance with manifest validation, API integration, and authentication issues.

How do I report a bug in an OpenAI plugin?

Open a GitHub issue at https://github.com/openai/plugins/issues and include the specific plugin name, reproduction steps, and any error logs. Reference the exact file path where the error occurred (such as plugins/zoom/skills/video-sdk/windows/examples/send-raw-video.md) and include your plugin.json configuration to help engineers reproduce the issue quickly.

Can I automate the inclusion of support contacts in my plugin?

Yes. Use the create_basic_plugin.py script located at .agents/skills/plugin-creator/scripts/create_basic_plugin.py. This helper automatically generates a plugin.json file with the email field pre-populated with [email protected], ensuring your plugin always includes the required support contact information.

Where is the plugin-creator skill configuration stored?

The configuration is stored in .agents/skills/plugin-creator/agents/openai.yaml. This file defines the OpenAI model that powers the skill and contains fallback contact entries pointing to the official support email, ensuring developers can always find help even when using the automated scaffolding tools.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}