Finalise Sprint Skill

This skill implements the COMPLETE + CRISP + LEARN framework to ensure professional sprint completion before shipping code.

Philosophy

"A sprint isn't done until the codebase tells the same story the code does." "Ship it clean, or ship it never."

When to Use

• After all sprint stories are complete (all tests passing)
• Before announcing/releasing the sprint deliverables
• After implementation phase but before commit/PR
• When CI/CD is green on the feature branch

Framework Overview

Framework	Focus	Checks
COMPLETE	Documentation & Tracking	8 checks
CRISP	Release Readiness & Quality	5 checks
LEARN	Knowledge Capture	5 checks

---

Part 1: COMPLETE (Documentation & Tracking)

C - CLAUDE.md Synchronisation (if applicable)

Check:

• Version number updated
• Last Updated date is current
• New features documented
• Quick Stats updated

O - Outstanding ADRs

CRITICAL: Ensure all architectural decisions are documented.

Check:

• Sprint ADR directory exists: ${sprint.base_dir}/adr/
• At least one ADR exists for major decisions
• ADRs follow the template format
• ADR README/index is updated

Common ADR Triggers:

• New external dependency added
• Major refactoring or architecture change
• New integration pattern
• Security-related changes
• API design decisions
• Database schema changes

ADR Template:

# ADR-NNNN: [Title]

**Status**: proposed | accepted | deprecated | superseded
**Date**: YYYY-MM-DD
**Sprint**: ${sprint_name}

## Context
What is the issue motivating this decision?

## Decision
What is the change we're making?

## Consequences

### Positive
- ...

### Negative
- ...

## Alternatives Considered
What other options were evaluated?

M - Master Tracking Files

Check:

• Sprint STORIES.md shows all stories complete
• Sprint README.md status is accurate
• No stories left as in_progress

P - Purge Artifacts

Check:

• No *.tmp, *.bak files
• No debug print statements in code
• No TODO/FIXME without tickets
• No large commented-out code blocks

L - Links and References

Check:

• Internal markdown links resolve
• Sprint docs reference each other correctly
• No broken links in documentation

E - External Documentation

Check:

• API documentation updated if APIs changed
• README updated if usage changed
• Migration guides created if breaking changes

T - Test Coverage

Check:

• All new code has tests
• Coverage meets threshold (90%+)
• No skipped tests without explanation
• Integration tests exist for new features

E - Epic Summary

Create: ${sprint.base_dir}/SUMMARY.md

# Sprint Summary: ${sprint_name}

**Completed**: [Date]
**Stories Delivered**: [Count]

## Key Achievements
1. [Achievement 1]
2. [Achievement 2]

## Technical Decisions (ADRs)
| ADR | Title | Status |
|-----|-------|--------|
| [N] | [Title] | accepted |

## Known Issues / Tech Debt
- [Issue 1]: [Impact]

## Lessons Learned
- [Lesson 1]

---

Part 2: CRISP (Release Readiness)

C - CI/CD Pipeline Verification

BLOCKING: Sprint cannot be complete if CI/CD is failing.

Check:

• All tests passing
• Build succeeds
• Type checking passes
• Linting passes (or acceptable warnings)

R - Repository Hygiene

Check:

• No uncommitted changes
• Feature branch up to date with main
• Clean git status

I - Integrity (Type Safety)

Check:

• mypy passes with no errors
• No new # type: ignore without justification
• Linting passes

S - Security Scan

Check:

• No hardcoded secrets or API keys
• No debug endpoints exposed
• .gitignore covers sensitive files

P - Packages

Check:

• No known security vulnerabilities
• Dependencies are current
• No unused dependencies

---

Part 3: LEARN (Knowledge Capture)

L - Lessons Captured

Document:

• What went well (keep doing)
• What could improve (do differently)
• What surprised us (didn't expect)
• Process improvements identified

E - Experience Documented

Check:

• Complex algorithms explained
• Non-obvious decisions documented
• Debugging tips captured

A - API Changes Noted

Check:

• Breaking changes documented
• New endpoints documented
• Deprecated endpoints marked

R - Release Notes

Create CHANGELOG entry:

## [Version] - [Date]

### Added
- [New feature]

### Changed
- [Change]

### Fixed
- [Bug fix]

N - Next Sprint Prepared

Check:

• Tech debt items logged
• Follow-up stories identified
• Handoff notes created

---

Output Artifacts

The finalise phase creates:

1. ${sprint.base_dir}/SUMMARY.md - Sprint completion summary
2. ${sprint.base_dir}/adr/ - Architecture Decision Records
3. Updated documentation - All docs current with implementation

Quality Gates

Gate	Requirement	Blocking?
Tests Pass	All green	Yes
ADR Exists	At least one	Yes
Summary Created	SUMMARY.md	Yes
Type Check	mypy passes	No
Lint Check	ruff/flake8	No

Interactive Checklist

When running finalise-sprint interactively:

1. Gather Information

   • Sprint name and directory
   • Epic numbers delivered

2. Run COMPLETE Checks

   • Documentation review
   • ADR verification
   • Artifact cleanup

3. Run CRISP Checks

   • CI/CD verification
   • Code quality checks
   • Security scan

4. Run LEARN Checks

   • Capture lessons
   • Document experience
   • Prepare handoff

5. Generate Report

   • Create SUMMARY.md
   • List all findings
   • Recommend actions

Example Usage

# Run finalise-sprint skill
adws run --phase finalise --sprint-name user-auth

# Dry run to see what would be checked
adws run --dry-run --phase finalise

Success Criteria

Sprint is ready to ship when:

• All COMPLETE checks pass
• All CRISP blocking checks pass
• LEARN documentation created
• SUMMARY.md exists
• At least one ADR exists
• Tests passing at 90%+ coverage

---

Related Skills

• /vision-creation - Create vision document
• /prd-creation - Create PRD from vision
• /architecture-design - Design system architecture
• /story-writing - Create user stories
• /code-review - Review code quality