mh-cto-sdlc-sprint-plan

Sprint Planning

Mode: synthesize-first (D34). Read the Phase 2 epic files and Phase 5 testing mandate. Derive the work breakdown and sequencing. Ask only where ordering priorities require a business decision not resolvable from the artifacts.

Elite character: A sprint plan that just lists all stories in creation order is not a plan — it is a transcript. Elite sprint planning sequences work by dependency, surfaces risk early, and produces a day-one-dev-ready state file that any engineer can pick up and begin from without a kickoff meeting.

Step 1 — Load Artifacts

Confirm present:

• Epic map (epic-map.md) ✓ / ✗ — REQUIRED
• Acceptance criteria (ac-*.md) ✓ / ✗ — REQUIRED
• Phase 5 gate verdict (phase5-gate-verdict.md) ✓ / ✗
• Phase 6 testing mandate (embedded in phase5-gate-verdict.md) ✓ / ✗
• Constraint inventory (constraint-inventory.md) ✓ / ✗

Missing epic-map.md or acceptance criteria: "Cannot build sprint plan without Phase 2 outputs. Ensure epic-map.md and ac-*.md exist."

Step 2 — Read and Inventory Epics

From the epic map, build a complete inventory of all epics and their stories:

For each epic:

• Epic ID and name
• Phase 2 acceptance criteria it delivers (link to ac-*.md)
• Story count and rough size (if story-spec.md exists)
• Dependencies (does this epic depend on another epic's outputs being in place?)
• Risk class: High / Medium / Low (based on tech complexity, external dependencies, unknown territory)

Step 3 — Sequence Epics and Stories

Apply sequencing rules:

Rule 1 — Dependency order: If Epic B requires Epic A's schema or API to exist, A must ship first. Map explicit data and API dependencies between epics.

Rule 2 — Risk-early: High-risk epics go first. Risk discovered in sprint 1 costs less than risk discovered in sprint 5. If a story contains novel technology or an integration not previously spiked, it goes early.

Rule 3 — Vertical slices (D23): Every story must be a complete end-to-end path through all layers. Stories that are one layer only (e.g., "add database schema") violate D23 and must be flagged as badly formed. Do not sequence them — flag them for revision before create-story runs.

Rule 4 — Foundation first within an epic: Within an epic, the story that establishes the foundational entity (the database schema + API + minimal UI) goes first. Subsequent stories extend it.

Produce a sequenced list:

Sprint 1:
  - [Story ID] {story name} — Epic: {epic} — Risk: {H/M/L} — Rationale: {why this order}
  - [Story ID] ...

Sprint 2:
  - ...

Sprints are advisory groupings, not commitments. The state machine is story-by-story. Sprint boundaries are a planning heuristic.

Step 4 — Flag D22/D23 Violations Before Writing State

Before writing the state file, flag any stories that violate the AGENT-BRIEF durability rules:

D22 violations (file paths or line numbers referenced in story text):

• Flag: "Story '{name}' in epic-map.md references a file path. This will rot within days of a refactor. Rephrase to reference the interface or module name instead."

D23 violations (horizontal slice — story covers only one layer):

• Flag: "Story '{name}' is a {schema-only / API-only / UI-only} slice. Every story must be end-to-end, independently demoable. Revise before this story enters the backlog."

Flagged stories are excluded from the state file until revised. Do not silently include them.

Step 5 — Write sprint-status.yaml

Save to _mh-output/cto/sdlc/sprint-status.yaml:

# Generated by mh-cto-sdlc-sprint-plan
# Last updated: {YYYY-MM-DD}
# Project: {project name from mh-project-context.md}

project: '{project name}'
phase: 6
generated_at: '{YYYY-MM-DD}'

epics:
  - id: '{epic-id}'
    name: '{epic name}'
    status: backlog # backlog | in-progress | done
    acceptance_criteria: ['{ac-id}', '{ac-id}']
    stories:
      - id: '{story-id}' # e.g., feat-auth-login-001
        name: '{story name}'
        epic: '{epic-id}'
        status: backlog # backlog | ready-for-dev | in-progress | review | done
        risk: high # high | medium | low
        sprint: 1 # advisory
        notes: '' # used by story-kickoff for session_state or flags
      - id: '{story-id}'
        name: '{story name}'
        epic: '{epic-id}'
        status: backlog
        risk: medium
        sprint: 1
        notes: ''

Status machine:

• Epics: backlog → in-progress → done
• Stories: backlog → ready-for-dev → in-progress → review → done

All stories start at backlog. create-story advances the first backlog story to ready-for-dev.

Step 6 — Write Output

Save to _mh-output/cto/sdlc/sprint-plan/{YYYY-MM-DD}_{r###}_sprint-plan.md:

# Sprint Plan

**Project:** {name}
**Date:** {date}
**Phase 5 Testing Mandate:** Carried forward (see phase5-gate-verdict.md)

---

## Epic Inventory

{Table: Epic ID | Name | Stories | Risk | Dependencies}

## Sequencing Rationale

{Explanation of why the sequence was chosen — top 3-5 decisions}

## Sprint Breakdown (advisory)

{Sprint-by-sprint story list with risk labels}

## D22/D23 Violations Flagged

{List of stories requiring revision before entering backlog — or "None"}

## Sprint-status.yaml Location

`_mh-output/cto/sdlc/sprint-status.yaml`

Step 7 — Present Summary

Sprint Plan complete.

Epics: {N} epics, {N} total stories across {N} advisory sprints.

Flags: {N} D22/D23 violations requiring story revision before dev begins.

High-risk stories (scheduled early): {list}

Next step: Run /mh-cto-sdlc-create-story to create the first development-ready story.