What are DataDoe Plugins?
Plugins give DataDoe agents reusable context and task-specific instructions. Instead of repeating the same background information or workflow in every conversation, add it once and let agents load it when relevant.
DataDoe MCP serves Plugins to every agent you connect, whatever platform that agent runs on, so you can work with several AI providers at the same time from one set of context. Enabled Plugins are available to agents; disabled ones are not.
How do I make my agents use Plugins?
DataDoe MCP exposes Plugins and encourages agents to use them. Agents behave differently, and some do not load Plugins until you ask for them.
Add the following short instruction to your agent-specific context, or at the end of your first prompt in the conversation: Use DataDoe Plugins.
For example:
How is my store performing this week?
Use DataDoe Plugins.What kinds of Plugins are there?
| Kind | Description | Example use case |
|---|---|---|
| DataDoe Skill | Ready-made playbooks maintained by DataDoe, installed in one step. | Add a weekly business review your agents can run straight away. |
| Memory | Something your agents should know or always follow, written by you. | "We sell private-label kitchen tools on Amazon US and EU." |
| Document | A file your agents can read when it is relevant. | Upload a PDF with brand guidelines. |
| Skill | One or more agent skills (opens in a new tab) from a public GitHub repository. | Add a custom or community reusable playbook to your agents. |
A Memory can carry an optional label: Business, Process, Brand, Metrics, Rules, Agent behaviour, or About me. A Document can carry Business or Agent behaviour. Labels help you find things later. They do not change how agents treat the Plugin.
What are Plugin scopes?
Plugins have two scopes:
- Organization Plugins are available to everyone in your Organization; they are useful for shared definitions, brand guidance, common business rules, and team workflows, and only Organization Owners can add or manage them
- Personal Plugins are yours alone; they are useful for individual preferences, responsibilities, and recurring ways of working
Use Organization Plugins for context and workflows that should be consistent across the team. Use Personal Plugins for context and workflows that apply only to your work.
What are the Plugin limits?
Each scope has its own allowance, counted separately:
- up to 10 active Plugins per scope, so 10 Organization Plugins and 10 Personal Plugins; every kind counts toward this
- a fixed amount of context space per scope, used only by Memories; the Plugins page shows how much of it you have used
- a single Memory holds up to 16,384 characters
Skills and Documents use no context space, so only the Plugin count applies to them. Plugins do not count toward your data row usage.
How to add a Memory?
- Open Plugins (opens in a new tab) from the side menu.
- Select Add organization plugin or Add personal plugin.
- Choose Memory.
- Give the Memory a name of up to 128 characters. Names accept letters, numbers, spaces, and dashes only.
- Add the information or instructions you want DataDoe agents to use, and optionally pick a label.
- Choose whether the Plugin should be enabled, then select Create.
How to add a Skill?
From the DataDoe library
- Open Plugins (opens in a new tab) from the side menu.
- Select Add organization plugin or Add personal plugin.
- Choose DataDoe Skill.
- Select the Skills you want, then select the install button.
From a GitHub repository
- Open Plugins (opens in a new tab) from the side menu.
- Select Add organization plugin or Add personal plugin.
- Choose Skill.
- Enter a public GitHub repository as
owner/repositoryor a GitHub URL. - Select Find skills.
- Choose up to 50 discovered Skills, then select the install button.
Install Skills only from repositories you trust. Review
SKILL.md, scripts, reference files, and assets before enabling a Skill.
DataDoe fetches Skill content when you install it and checks the repository for updates roughly every three days. Later changes are picked up automatically.
How to add a Document?
- Open Plugins (opens in a new tab) from the side menu.
- Select Add organization plugin or Add personal plugin.
- Choose Document.
- Add a description of up to 2,048 characters covering what the file contains and why agents should use it.
- Upload a supported file of up to 2 MB, then select Create.
DataDoe converts the file to markdown and keeps the converted content with the Plugin.
The description decides whether the file is ever used. Agents read it to choose which Documents to load, so state what the file contains and when it applies.
Supported formats: pdf, doc, docx, ppt, pptx, xls, xlsx, csv, html, htm, xml, json, epub, zip, txt, md, markdown, msg, png, jpg, jpeg, gif, webp.
Plugins in MCP
Agents work with Plugins through these DataDoe MCP tools:
| Tool | Purpose |
|---|---|
plugins_get | List the enabled Plugins for the user |
plugins_skills_get | Get a Skill's instructions or one of its supporting files |
plugins_files_get | Get the converted content of a Document |
plugins_memories_create | Create a Memory |
plugins_memories_edit | Change the content of a Memory |
The last two write. Agents can suggest and then create or update Memories themselves, so a Memory can change without you opening the Plugins page.
Plugins need an MCP key bound to your user. If your agent reports that the key is missing a personal user binding, regenerate the key in Integrations (opens in a new tab).
Tips for effective Plugins
- Keep each Plugin focused on one topic.
- Use a clear, descriptive name.
- Write instructions in direct language and include examples where helpful.
- Prefer focused documents over large archives, and upload only files you trust.
- Review Organization Plugins regularly so shared guidance stays accurate.
- Disable outdated Plugins instead of leaving conflicting instructions active.
- Do not store passwords, access keys, or other secrets in Plugins.
Related documentation
- Read the Features Overview for other DataDoe features.
- Read the DataDoe MCP overview for connecting an AI agent to DataDoe.

