Connecting the NimbleWork MCP to Codex (CLI) (PDF)<\/u><\/a><\/li>\n<\/ul>\nWhat You Can Do<\/h2>\n
The NimbleWork connector works with your Nimble Workitems (cards)<\/strong> and the items on them \u2014 it does not cover everything you can do in Nimble. Once connected, you can use natural language to:<\/p>\n\n- Find and read<\/strong> \u2014 list your projects and workspaces and their card types; read Workitems and their ToDos, comments, attachments, and linked Workitems.<\/li>\n
- Create and update Workitems<\/strong> \u2014 create a Workitem, update its details, and close, reopen, or delete it.<\/li>\n
- Manage ToDos and time<\/strong> \u2014 add ToDos to a Workitem, assign and update them, close, reopen, or delete them, and log time.<\/li>\n
- Act on Workitems<\/strong> \u2014 flag, block, or vote on a Workitem (and undo each); link or unlink Workitems (for example, as parent and child); add a comment; or attach a file.<\/li>\n
- Switch context<\/strong> \u2014 select the project or workspace to work in.<\/li>\n<\/ul>\n
Tip:<\/strong> If your prompt does not specify a workspace and you have access to more than one, the AI assistant will ask you to choose before taking any action \u2014 so your changes always land in the right place.<\/p>\nYour Access and Permissions<\/h2>\n
The NimbleWork connector works with exactly the access you already have inside Nimble \u2014 no more, no less. Signing in through your AI assistant never gives you extra rights. If you can do something in Nimble, you can do it through the connector; if you can’t, the connector can’t either.<\/p>\n
This is because Nimble decides what you can see and do based on your persona<\/strong> (your assigned role) and the projects you’ve been added to. The connector simply carries those same rules across to your AI assistant. (You may hear this called role-based access control<\/em>.)<\/p>\nWhat this means for you<\/strong><\/p>\n\n- You’ll only see the projects and workspaces you’ve been invited to \u2014 not every project in the account.<\/li>\n
- You can only work with the workitem types your persona is allowed to use.<\/li>\n
- For each workitem type, you can do only the actions your persona permits \u2014 for example view only<\/em>, view and edit<\/em>, or full access<\/em>.<\/li>\n
- If you try something you’re not allowed to do, the assistant tells you it couldn’t be done because you don’t have permission \u2014 and nothing changes in Nimble.<\/li>\n<\/ul>\n
Example \u2014 what to expect<\/strong><\/p>\nSay your persona is a Reviewer<\/strong> who can view everything in a project, create and edit User Stories, but not delete them \u2014 and can only view Releases.<\/p>\n\n- “Show me the user stories in Project Apollo.” \u2192 The assistant lists them. \u2705<\/li>\n
- “Create a user story called ‘Login screen redesign’.” \u2192 It’s created. \u2705<\/li>\n
- “Delete user story US-42.” \u2192 The assistant replies that you don’t have permission for this action; the story stays untouched. \u274c<\/li>\n
- “Change the dates on Release 3.” \u2192 You can view the release but not edit it, so the assistant reports it can’t make that change. \u274c<\/li>\n<\/ul>\n
In short, the connector is a safe way to work: it can never let anyone do something their Nimble role doesn’t already allow.<\/p>\n
The connector exposes the same set of tools on every supported AI platform \u2014 31 in total, grouped into read-only tools and write\/delete tools. Read-only tools only retrieve information; write\/delete tools make changes. Most AI assistants ask for your approval before a tool runs, and typically let you allow an action once or always allow it.<\/p>\n
Read-only tools (9)<\/strong> \u2014 these only retrieve information; they never change anything.<\/p>\n\n\n\nTool<\/strong><\/td>\nWhat it does<\/strong><\/td>\n<\/tr>\n\n| List Workspaces<\/td>\n | Shows the projects and workspaces your account can access.<\/td>\n<\/tr>\n | \n| List Forms<\/td>\n | Lists the workitem types (forms) available in a workspace.<\/td>\n<\/tr>\n | \n| List User Forms<\/td>\n | Lists the workitem types you personally have access to.<\/td>\n<\/tr>\n | \n| Get Form<\/td>\n | Shows the fields and structure of a specific workitem type.<\/td>\n<\/tr>\n | \n| List Cards<\/td>\n | Lists workitems in a workspace, with optional filters (by type, status, and more).<\/td>\n<\/tr>\n | \n| Get Linked Cards<\/td>\n | Shows the workitems linked to a given workitem \u2014 its parents, children, and related items.<\/td>\n<\/tr>\n | \n| List Todos<\/td>\n | Lists the ToDos on a workitem.<\/td>\n<\/tr>\n | \n| Get Comments<\/td>\n | Retrieves all comments added to a workitem.<\/td>\n<\/tr>\n | \n| Get Attachments<\/td>\n | Retrieves the details of the attachments on a workitem (such as name and file size), not the attachment files themselves.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n Write \/ delete tools (22)<\/strong> \u2014 these make changes, and always respect your Nimble permissions.<\/p>\n\n\n\nTool<\/strong><\/td>\nWhat it does<\/strong><\/td>\n<\/tr>\n\n| Create Card<\/td>\n | Creates a new workitem.<\/td>\n<\/tr>\n | \n| Update Card<\/td>\n | Edits fields on an existing workitem (name, description, dates, assignee, and more).<\/td>\n<\/tr>\n | \n| Close Card<\/td>\n | Marks a workitem as complete\/closed.<\/td>\n<\/tr>\n | \n| Reopen Card<\/td>\n | Reopens a closed workitem.<\/td>\n<\/tr>\n | \n| Delete Card<\/td>\n | Deletes a workitem.<\/td>\n<\/tr>\n | \n| Flag Card<\/td>\n | Flags a workitem to draw attention to it.<\/td>\n<\/tr>\n | \n| Unflag Card<\/td>\n | Removes a flag from a workitem.<\/td>\n<\/tr>\n | \n| Block Card<\/td>\n | Marks a workitem as blocked, with a reason.<\/td>\n<\/tr>\n | \n| Unblock Card<\/td>\n | Removes the blocked status from a workitem.<\/td>\n<\/tr>\n | \n| Vote on Card<\/td>\n | Adds your vote to a workitem.<\/td>\n<\/tr>\n | \n| Remove Vote from Card<\/td>\n | Removes your vote from a workitem.<\/td>\n<\/tr>\n | \n| Link Card<\/td>\n | Links one workitem to another (for example, parent, child, or related).<\/td>\n<\/tr>\n | \n| Unlink Card<\/td>\n | Removes a link between two workitems.<\/td>\n<\/tr>\n | \n| Add Todo<\/td>\n | Adds a ToDo (checklist item) to a workitem.<\/td>\n<\/tr>\n | \n| Update Todo<\/td>\n | Edits a ToDo \u2014 for example its text, assignee, or estimate.<\/td>\n<\/tr>\n | \n| Close Todo<\/td>\n | Marks a ToDo as done.<\/td>\n<\/tr>\n | \n| Reopen Todo<\/td>\n | Reopens a completed ToDo.<\/td>\n<\/tr>\n | \n| Delete Todo<\/td>\n | Removes a ToDo from a workitem.<\/td>\n<\/tr>\n | \n| Log Time<\/td>\n | Records time spent against a workitem or its ToDo.<\/td>\n<\/tr>\n | \n| Add Comment<\/td>\n | Posts a comment on a workitem. Note:<\/strong> @mentions and emoji are added as plain text through the connector \u2014 they do not trigger notifications or render the way they do in the product UI.<\/td>\n<\/tr>\n\n| Add Attachment<\/td>\n | Attaches a file to a workitem. Note:<\/strong> Only small, text-based files (up to a few KB, such as TXT or CSV) can be attached through the connector. For larger files, such as multi-MB PDFs or images, add the attachment from the product UI in the Comments & Attachments tab.<\/td>\n<\/tr>\n\n| Select Workspace<\/td>\n | Sets the active project\/workspace so your next requests act in the right place.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\nExample Actions<\/h2>\nThe examples below show natural language prompts you can type and what the connector does in response.<\/p>\n 1. List your workspaces<\/h3>\nPrompt:<\/strong> \u201cList all the workspaces I have access to.\u201d<\/p>\nWhat happens:<\/strong> The connector returns the names and IDs of all workspaces accessible to your account. Use this to confirm which workspace you want to work in before creating or updating workitems.<\/p>\n2. Create a workitem<\/h3>\nPrompt:<\/strong> \u201cCreate a user story in the Feature Development project named ‘Design for Mobile login page’.\u201d<\/p>\nWhat happens:<\/strong> The connector creates a new card in the specified workspace and form, and returns its ID (e.g., FF21) and a confirmation. You can include additional fields such as priority, description, estimate, or due date in your prompt.<\/p>\n3. Update a workitem<\/h3>\nPrompt:<\/strong> \u201cUpdate FF21 \u2014 set the description to describe the mobile login mockups and set the estimate to 6 hours.\u201d<\/p>\nWhat happens:<\/strong> The connector updates only the fields you specify (description and EstimateHours) on FF21. All other fields remain unchanged.<\/p>\n4. Add a todo and log time<\/h3>\nPrompt:<\/strong> \u201cAdd a todo to FF21 called ‘Create wireframes’ with an estimate of 3 hours, then log 2 hours of work against it.\u201d<\/p>\nWhat happens:<\/strong> The connector creates the todo on FF21 and logs 2 hours of actual effort against it in a single conversation \u2014 no switching between tools required.<\/p>\n | | | | | | | |
![\"Using](\"https:\/\/digitecdn.s3.us-east-1.amazonaws.com\/images\/nimblework_mcp_10.0.gif\")
<\/p>\n