Sprint Planning Session

Sprint Planning Session — Entry Orchestrator

NON-NEGOTIABLE (read first)

1. Save every artifact to file before presenting in chat. Chat is ephemeral.
2. Never fabricate dates, numbers, statuses, or capacity. Unknown = [TBD]. Inferred = [INFERRED].
3. Never skip checkpoints. Every CHECKPOINT requires user confirmation before proceeding.
4. Graded provenance: authoritative claims (dates, capacity numbers, statuses, decisions, constraints) require a source tag. Synthesized analysis uses (Synthesized from: ...). Process scaffolding (section framing, transitions) is exempt. Missing source on an authoritative claim = flag as (Source: Implicit) for resolution.
5. One stage at a time. Complete and save before starting the next.
6. Re-read upstream documents at the start of every workflow. Never rely on cached context.
7. Every ticket in the input must appear in the output. None silently dropped.

Purpose

Take the sprint's planned work (from the release plan + ticket list) and produce a document that gives the team clarity on what they're building, why, and what "done" looks like.

What you'll produce:

• [sprint-root]/Sprint-[N]-Planning.md — the sprint planning session document

State file: [release-root]/.meta/skill-state-sprint-planning-session.json — tracks current task and artifacts for resume capability. Updated after each workflow step.

When to invoke: Before every sprint kickoff. The PM has tickets in their board (ADO, Jira, or a list) and needs to turn them into a structured planning document.

Workflow Chain

[You are here]
SKILL.md (Intake)
     ↓ reads
workflows/01-intake.md        → normalizes tickets, detects carryover
     ↓ reads
workflows/02-organize.md      → groups tickets, runs guardrails
     ↓ reads
workflows/03-generate.md      → produces Sprint-N-Planning.md
     ↓ reads (optional)
workflows/04-refine.md        → applies user feedback

Each workflow file is read in turn. Do NOT skip ahead.

Step 1: Determine Release Context

Ask the user:

1. Release root folder — Where are the release artifacts? (e.g., Product Artifacts/Delivery Plan/Releases/Redesign_Aug2026/)
2. Sprint number — Which sprint is this?

If the user provides both, scan the release root for:

• Release-Plan.md — for sprint goal and planned scope
• Sprint-[N-1]/Sprint-[N-1]-Review.md or Sprint-[N-1]-Progress.md — for carryover detection
• .meta/constraint-registry.md — for active constraints

Report what you found:

Found in release root:
- Release-Plan.md: ✅ Found / ❌ Not found
- Prior sprint review: ✅ Found / ❌ Not found
- Constraint registry: ✅ Found / ❌ Not found

If no release root is provided, that's fine — the skill works standalone with just a ticket list.

Step 2: Gather Sprint Inputs

Ask the user for:

• Sprint dates (start and end) — or infer from Release Plan if linked
• Sprint goal — or extract from Release Plan if linked
• Ticket list — accept ANY format: paste from ADO/Jira, CSV, markdown table, verbal list, screenshot
• Team capacity — who is available, any PTO or reduced capacity

Step 3: Read Shared Context

MANDATORY READ the following files before proceeding:

Read the file:

../shared/delivery-model.md

Extract the sprint anatomy and commitment model. Understand where this sprint sits in the release lifecycle.

Read the file:

../shared/execution-rules.md

Load the shared NON-NEGOTIABLE rules, quality criteria, self-check structure, and provenance model.

If a constraint registry was found in Step 1, read it and note active constraints that affect this sprint.

Step 4: Start Intake Workflow

Read the file:

workflows/01-intake.md

Follow that file's instructions completely from start to finish. When complete, return here and proceed to Step 5.

Step 5: Organize and Validate

Read the file:

workflows/02-organize.md

Follow that file's instructions completely. When complete, return here and proceed to Step 6.

Step 6: Generate Planning Document

Read the file:

workflows/03-generate.md

Follow that file's instructions completely. When complete, return here and proceed to Step 7.

Step 7: Refine (if needed)

If the user requests changes after reviewing the generated document:

Read the file:

workflows/04-refine.md

Otherwise, the skill is complete.

Related Skills