Sprint Complete

Sprint Complete

Instructions

Complete the current sprint by validating delivery, running the E2E regression gate, closing GitHub issues, generating a summary report, updating the project dashboard, archiving sprint artifacts, and creating a pull request to merge the sprint branch into main.

The PR serves as the formal release gate — E2E tests run in GitHub Actions on the PR, and a formal code review must be completed before merging. Merging the PR to main is the release / sprint demo moment.

Prerequisites

• Sprint has been delivered (/sprint-deliver completed for all use cases in scope)
• The application builds successfully (npx next build)
• docs/sprints/next-sprint/readiness-report.md exists
• Currently on a sprint- branch (created by /sprint-kickoff)

If any prerequisite fails, stop and report which prerequisite is not met.

DO NOT

• Modify specification or design documents
• Skip the E2E regression gate when E2E tests exist
• Fix bugs or implementation issues — the sprint is done, this is closing time
• Skip the GitHub issue closure step
• Archive without generating the summary first
• Create a git commit on the current branch without user confirmation
• Create the PR without user confirmation
• Sleep or wait between test retries — when npx playwright test fails, present options to the user immediately. Never use sleep, setTimeout, or any delay before re-running. The fix-then-rerun cycle must be immediate — no pauses of any duration

Nexa Rules Gate

Read and follow ${CLAUDE_PLUGIN_ROOT}/shared/readiness/NEXA_RULES_GATE.md.

Pipeline

Execute these phases in order.

---

Phase 1: Validate Sprint Completion

1. Check that docs/sprints/next-sprint/ exists.

• If not, stop: "No active sprint found. Run /sprint-prepare to start a sprint."

1. Read docs/sprints/next-sprint/readiness-report.md to get the sprint scope.

• Extract the Delivery Order table to know which UCs were planned.

1. Check delivery status for each planned UC:

• Look for docs/delivery/-iterations.md
• Mark as delivered if the iteration log exists

1. If any UCs are not delivered, present the status and ask the user:

> Sprint has undelivered use cases: > > | UC ID | Name | Status | > |-------|------|--------| > | UC-XXX | [name] | Delivered | > | UC-YYY | [name] | NOT DELIVERED | > > Options: > 1. Complete anyway — undelivered UCs will be noted in the summary and excluded > from GitHub issue closure > 2. Cancel and continue delivery > > What would you like to do?

Wait for user confirmation. If the user cancels, stop.

---

Phase 2: E2E Regression Gate

Run the full Playwright E2E suite as a regression check before closing the sprint. This ensures that bug fixes, technical tasks, and any other changes introduced during the sprint have not broken existing functionality — even in sprints that only contain bug fixes and have no new use case deliveries.

1. Check whether E2E tests exist:

• Look for files matching e2e/**/*.spec.ts
• If no test files exist, skip this phase entirely and note "No E2E tests found — regression

gate skipped" (this will be included in the sprint summary)

1. Verify Docker is running (required for Testcontainers database):

``bash docker info > /dev/null 2>&1 `` If Docker is not running, ask the user to start Docker or Colima before continuing. Wait for confirmation, then re-check before proceeding.

1. Run the full E2E suite:

``bash npx playwright test ` **Critical rules** (same as /deliver-use-case`):

• NO filters: no --grep, --grep-invert, --project subset
• Must show 0 failed, 0 skipped, exit code 0 to pass

1. If all tests pass:

• Record the result: total tests, passed count, duration
• Proceed to Phase 3

1. If any tests fail:

• Display the failure summary to the user
• Classify each failure:
• Regression from bug fix — a fix introduced in this sprint broke an existing test
• Flaky test — test passes on re-run without code changes
• Pre-existing failure — test was already failing before this sprint
• Present options:

> E2E Regression Gate — FAILED > > | Test | Classification | Details | > |------|---------------|---------| > | [test name] | [type] | [brief reason] | > > Options: > 1. Fix and re-run — address the failures, then re-run the suite > 2. Complete anyway — proceed with failures noted in the sprint summary > 3. Cancel — stop sprint completion > > What would you like to do?

• Option 1: The user fixes the issue (or asks you to), then re-run from step 3.

Allow up to 3 re-runs. After 3 failures, only options 2 and 3 remain.

• Option 2: Record failures in the sprint summary as known regressions and proceed.
• Option 3: Stop the pipeline.

---

Phase 3: Close GitHub Issues

For each delivered use case, find and close its GitHub issue:

1. Search for the issue:

`` gh issue list --search "in:title " --state open --json number,title -q '.[] | select(.title | startswith(":")) | .number' ``

1. If an open issue is found, close it with a comment:

`` gh issue close --comment "Delivered in sprint-N. Closing as part of /sprint-complete." ``

1. Track results:

| UC ID | Issue | Action | |-------|-------|--------| | UC-XXX | #42 | Closed | | UC-YYY | — | No issue found |

Report the results to the user before proceeding.

---

Phase 4: Determine Sprint Number

1. Read the sprint number from the current branch name:

``bash git branch --show-current ``

• Expected format: sprint- (e.g., sprint-3)
• Extract N from the branch name
• If the current branch does not match sprint-, stop and report:

"Not on a sprint branch. Expected sprint- branch (created by /sprint-kickoff)."

---

Phase 5: Generate Sprint Summary

Gather information and generate the summary:

1. Delivered UCs: List from iteration logs with iteration counts
2. Traceability: Read docs/delivery/-traceability.md for each delivered UC

to count verified requirements

1. E2E regression: Record result from Phase 2 (passed/failed/skipped, test counts, failures)
2. Issues encountered: Extract from iteration logs (retries, blockers, failures)
3. Sprint duration: From git log — first and last commit touching sprint artifacts
4. Audit results: Read docs/delivery/sprint-audit-report.md if it exists (optional —

may not be present since sprint-audit is no longer part of the pipeline)

Write docs/sprints/next-sprint/SUMMARY.md:

# Sprint N Summary

| | |
|---|---|
| **Completed** | [date] |
| **Duration** | [first delivery commit] → [last delivery commit] |

## Delivered Use Cases

| # | UC ID | Name | Iterations | Verified FRs | Status |
|---|-------|------|------------|--------------|--------|
| 1 | UC-XXX | [name] | N | N/N | Delivered |
| 2 | UC-YYY | [name] | N | N/N | Delivered |
| 3 | UC-ZZZ | [name] | — | — | Not delivered |

## Sprint Metrics

| Metric | Value |
|--------|-------|
| Planned UCs | N |
| Delivered UCs | N |
| Completion rate | N% |
| Total delivery iterations | N |
| Avg iterations per UC | N.N |
| Total verified FRs | N |

## E2E Regression Gate

| Metric | Value |
|--------|-------|
| Result | PASSED / FAILED (with known regressions) / SKIPPED (no tests) |
| Total tests | N |
| Passed | N |
| Failed | N |
| Duration | Ns |

*(If failed with known regressions, list them here:)*
- [test name] — [classification and brief reason]

## GitHub Issues

| UC ID | Issue | Action |
|-------|-------|--------|
| UC-XXX | #42 | Closed |
| UC-YYY | #43 | Closed |

## Issues Encountered

### Retries

- UC-XXX required N retries due to [reason from iteration log]

### Blockers

- [Any blockers encountered and how they were resolved]

*(If none: "No significant issues encountered.")*

## Carryover

- [ ] UC-ZZZ — not delivered, reason: [from Phase 1]
- [ ] [Any technical debt noted during delivery]

*(If none: "Nothing to carry over.")*

---

*Generated by `/sprint-complete`*

---

Phase 6: Update Sprints Overview Dashboard

The sprints overview lives at docs/sprints/sprints-overview/ and is driven by manifest.json. /sprint-prepare bootstraps this directory and creates the initial manifest entry with status in-progress.

1. Read docs/sprints/sprints-overview/manifest.json.

1. Find the sprint entry where status is "in-progress" (there should be exactly one —

the sprint being completed).

1. Update the entry:

• Set status to "delivered"
• Set date to today's date
• Update the files object to point to the archived sprint folder paths

(since the folder will be renamed in Phase 7): ``json "files": { "readiness": "../sprint-N/readiness-report.md", "changelog": "../sprint-N/changelog.md", "refinement": "../sprint-N/refinement-proposal.md", "summary": "../sprint-N/SUMMARY.md" } ``

1. Update currentSprint to null (no active sprint after completion).

1. Verify that the useCases array in the manifest contains entries for all delivered UCs

with correct paths to their spec, design, and delivery files. These entries should already exist (created by /sprint-prepare). If any are missing, add them.

1. Write the updated manifest.json.

---

Phase 7: Archive the Sprint

1. Rename the sprint folder:

``bash mv docs/sprints/next-sprint docs/sprints/sprint-N ``

1. Move delivery artifacts into the sprint archive:

```bash mkdir -p docs/sprints/sprint-N/delivery

# Move iteration logs for delivered UCs for each delivered UC-ID: mv docs/delivery/-iterations.md docs/sprints/sprint-N/delivery/

# Move traceability reports for delivered UCs for each delivered UC-ID: mv docs/delivery/-traceability.md docs/sprints/sprint-N/delivery/ 2>/dev/null || true ```

1. Create fresh docs/sprints/next-sprint/ directory:

``bash mkdir -p docs/sprints/next-sprint ``

---

Phase 8: Commit Sprint Artifacts

1. Stage all changes on the sprint branch:

``bash git add docs/sprints/ docs/delivery/ ``

1. Ask the user for confirmation before committing:

> Ready to commit sprint completion artifacts. > > Changes: > - Sprint archived to docs/sprints/sprint-N/ > - Delivery logs moved to archive > - Dashboard updated (manifest.json) > - Sprint summary generated > - Fresh docs/sprints/next-sprint/ created > > Create commit?

1. If confirmed, commit:

``bash git commit -m "chore(sprint-N): close — [1-line summary of delivered UCs]" ``

1. Push to remote:

``bash git push origin sprint- ``

---

Phase 9: Create Pull Request

Create a PR from sprint- → main. The sprint report becomes the PR body so that reviewers see the full sprint summary, and GitHub Actions runs E2E tests on the PR.

1. Read the generated docs/sprints/sprint-N/SUMMARY.md to use as the PR body content.

1. Ask the user for confirmation:

> Ready to create the sprint PR. > > - From: sprint- > - To: main > - Title: Sprint N: [1-line summary] > - Body: Sprint summary report > > Create the pull request?

1. If confirmed, create the PR:

``bash gh pr create --base main --head sprint- \ --title "Sprint N: [1-line summary of delivered UCs]" \ --body "" ``

1. Report to user:

> Sprint N completed! > > Delivered: > - UC-XXX: [name] > - UC-YYY: [name] > > GitHub Issues Closed: #42, #43 > > Pull Request: [PR URL] > > Dashboard: docs/sprints/sprints-overview/index.html > > Summary: docs/sprints/sprint-N/SUMMARY.md > > Next steps: > 1. GitHub Actions will run E2E tests on the PR > 2. Perform a formal code review on the PR > 3. Merge the PR to main — this is the release > 4. Present the dashboard to stakeholders for sprint demo > 5. Run /sprint-prepare to plan the next sprint

Folder Structure After Completion

docs/
├── sprints/
│   ├── sprints-overview/
│   │   ├── index.html          # Dashboard (cumulative, all sprints)
│   │   ├── manifest.json       # Updated: sprint-N status = delivered
│   │   └── md-viewer.html      # Markdown renderer
│   ├── sprint-1/
│   │   ├── readiness-report.md
│   │   ├── changelog.md
│   │   ├── refinement-proposal.md
│   │   ├── SUMMARY.md
│   │   └── delivery/
│   │       ├── UC-001-iterations.md
│   │       └── UC-001-traceability.md
│   ├── sprint-2/
│   │   ├── ...
│   └── next-sprint/             # Fresh, empty, ready for next sprint
├── delivery/                    # Cleared of completed sprint's artifacts
├── use_cases/                   # Specs remain in place (referenced by dashboard)
└── designs/                     # Designs remain in place (referenced by dashboard)

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: nexadevapp
• Source: nexadevapp/nexa-claude-skills-marketplace
• License: Apache-2.0

Install and usage instructions live in the source repository linked above.