Doing Real Work with the DevStride MCP Server

Once your AI client is connected, you talk to your work in plain language. This page is a library of prompts that work well, organized by what you're trying to do — plus a worked example of the assistant planning and sequencing a sprint end to end.

Get oriented

Before diving in, ask your assistant what it can do:

• "What tools are available from the DevStride MCP?" — see what's connected and understand its capabilities.
• "What can you do with these DevStride tools?"

Personal productivity

• "What open items are assigned to me?" — see your active work without opening a board.
• "Summarize my highest-priority work for this week."
• "Which of my items are blocked or at risk?"
• "Look at every item assigned to me with activity in the last 60 days. Read each description and suggest where I should make it clearer or more thorough — cross-reference the linked pull request when one exists."

Team & delivery visibility

• "Summarize the current risks across the board name board." — great for standups and reviews.
• "What work hasn't been updated recently?" — surface stale or abandoned items.
• "Show overdue items grouped by assignee."

Reporting & communication

• "Draft a status update for leadership based on recent activity."
• "Summarize what changed this week on the roadmap."

Process & workflow insights

• "Which work items are creating bottlenecks in delivery?"
• "Show work items missing estimates, owners, or target dates." — improve planning quality.

Advanced prompts

For power users and technical teams:

• "Find dependencies affecting release readiness."
• "Summarize cross-team delivery risks."
• "Show items related to theme, e.g. API modernization."
• "Generate release notes from completed work."
• "Identify trends in blocked work over the last 30 days."
• "Compare planned vs. completed work this sprint."

A worked example: plan and sequence a sprint

Here's the kind of multi-step work the MCP makes possible. Because the assistant works with roadmaps, initiatives, estimates, dates, and dependencies as real objects, it can carry a planning task from analysis all the way to a finished Gantt — pausing for your approval before it changes anything.

A typical flow:

1. Estimate a backlog. "Read the descriptions of every item in the New lane and propose a time estimate for each. Wait for my approval before changing anything." The assistant reads each item (and any linked pull request) and returns a table of proposed estimates.
2. Sequence the work. It identifies dependencies between items — what must happen first — and proposes an order, flagging the largest items and any risks.
3. Strengthen thin items. Where a description is too sparse to act on, it proposes an expanded version — objective, detail, and acceptance criteria — and a revised estimate, for you to approve or edit.
4. Apply the plan. Once you approve, it updates each item's estimate and sets start and due dates per the schedule.
5. Build the roadmap. It groups the items into initiatives by theme and creates the roadmap (a Gantt timeline) with those initiatives in DevStride.

The result is a fully structured, dated roadmap built from your real backlog — drafted by the assistant, but reviewed and approved by you at each step.

Working with sub-tasks

Your assistant can search across sub-tasks, list the sub-tasks of a parent, and create, update, or delete them. Item searches can also include items whose sub-task is assigned to you — so you can ask for "everything I'm working on, even via sub-tasks." Results are always filtered to what you have permission to read.