Sprint planning

Sprint Planning

Scrum Guide grounding

Sprint Planning kicks off the Sprint. The whole Scrum Team attends — Developers, the Product Owner, and the Scrum Master. It is time-boxed to eight hours for a one-month Sprint (shorter for shorter Sprints).

The event answers three questions:

Topic	Question
Why	What is the Sprint Goal — the single objective that makes this Sprint worthwhile?
What	Which Product Backlog items can the Developers commit to completing?
How	How will the selected work get done?

The Sprint Goal is the output that matters most. It gives the team coherence and the flexibility to negotiate scope during the Sprint.

---

Tool detection

Identify which project management tool is available before proceeding:

1. Check for active mcp__azure-devops__* tools → set $PM_TOOL to ado
2. Otherwise check for active mcp__jira__* tools → set $PM_TOOL to jira
3. If both → ask: "I see both ADO and Jira connected — which should I use?"
4. If neither → set $PM_TOOL to manual

---

Step 1 — Establish context

Ask for any missing context before pulling data:

• "What is the Sprint length?" (default: 2 weeks)
• "What is the team's capacity this Sprint?" (total developer-days, accounting for time off)
• "Is there a Product Goal or theme the Sprint should contribute to?"

Store as $SPRINT_LENGTH, $CAPACITY, $PRODUCT_GOAL.

---

Step 2 — Fetch the top of the Product Backlog

Retrieve the ordered, refined items at the top of the backlog:

• ADO: use wit_list_backlog_work_items or wit_list_backlogs for the current project
• Jira: use the equivalent backlog listing tool, ordered by rank
• Manual: ask the Product Owner to paste or describe the top 10–15 items

Present the items as a numbered list with ID, title, and current state. Flag any item that appears unrefined (no acceptance criteria, no size estimate) with a ⚠️.

---

Step 3 — Craft the Sprint Goal

A good Sprint Goal is:

• A single, coherent objective — not a list of features
• Outcome-oriented, not a task list
• Achievable within the Sprint
• Meaningful to stakeholders

Suggest a draft Sprint Goal based on the top backlog items and the Product Goal. Examples of good framing:

• "Enable customers to complete checkout without a phone call to support."
• "Reduce report generation time so finance can close the books faster."

Ask: "Does this capture what this Sprint is for? Adjust it until it feels right."

Store the confirmed goal as $SPRINT_GOAL.

---

Step 4 — Select the Sprint Backlog

Guide the team to select items that:

1. Directly serve the Sprint Goal
2. Fit within $CAPACITY
3. Meet the Definition of Ready (refined, sized, unblocked)

Present a proposed selection. For each item, note:

• Whether it contributes to the Sprint Goal (✅) or is additional scope (📌)
• Any refinement gaps that should be addressed before committing

If the team wants to add items beyond what capacity allows, name the trade-off explicitly: "Adding #1234 would fill remaining capacity — is that the right call?"

---

Step 5 — Confirm and create the Sprint

Once the team confirms the Sprint Goal and backlog selection:

Present a summary:

Sprint Goal: [SPRINT_GOAL]

Committed items ([N] total):
- [#ID] [Title] — [contributes to goal / additional scope]
- ...

Items deferred to next Sprint:
- [#ID] [Title]

Ask: "Does this reflect the team's plan? Say 'looks good' to finalize, or tell me what to adjust."

After confirmation:

• ADO: use work_create_iterations or work_assign_iterations to create/assign the sprint; use wit_update_work_item to move items into the iteration
• Jira: use the sprint creation and issue-assignment tools
• Manual: output the summary in a clean block for the team to act on

---

Guardrails

• Never set the Sprint Goal on behalf of the Product Owner without confirmation.
• Never move items into a sprint without explicit team confirmation.
• Never estimate story points or size — that belongs to the Developers.
• If a proposed Sprint Goal is actually a list of features, coach gently: "A Sprint Goal is usually a single outcome. Would [X] work as the unifying purpose?"
• If capacity is unknown, proceed but flag that commitment is approximate.
• Never schedule work beyond the Sprint length — call out over-commitment explicitly.