feat: add night-watch summary command for morning briefings

[jonit-dev]

Summary

Adds a new night-watch summary command that provides a "morning briefing" aggregating data from multiple sources into a single view:

• Recent job runs from getJobRunsAnalytics()
• Open PRs from collectPrInfo()
• Pending queue items from getQueueStatus()

Usage

night-watch summary [--hours <n>] [--json]

Options

• --hours <n>: Time window in hours (default: 12)
• --json: Output summary as JSON

Examples

night-watch summary              # Last 12 hours
night-watch summary --hours 24   # Last 24 hours  
night-watch summary --json       # JSON output for scripting

Output Format

Night Watch Summary (last 12h)
────────────────────────────────

Jobs Executed: 5
  ✓ 3 succeeded   ✗ 1 failed   ⏱ 1 timed out

┌──────────┬──────────┬──────────┬──────────┬──────────────┐
│ Job      │ Status   │ Project  │ Provider │ Duration     │
├──────────┼──────────┼──────────┼──────────┼──────────────┤
│ executor │ success  │ my-app   │ claude   │ 8m 32s       │
│ reviewer │ success  │ my-app   │ codex    │ 4m 15s       │
│ executor │ failure  │ my-lib   │ claude   │ 12m 01s      │
└──────────┴──────────┴──────────┴──────────┴──────────────┘

Open PRs (2)
┌────┬────────────────────────┬──────────┬───────┐
│ #  │ Title                  │ CI       │ Score │
├────┼────────────────────────┼──────────┼───────┤
│ 42 │ feat: add login page   │ pass     │ 85    │
│ 43 │ fix: memory leak       │ pending  │ -     │
└────┴────────────────────────┴──────────┴───────┘

Queue: 1 pending (executor for my-lib)

No action needed — all jobs healthy.

When there are failures:

⚠ Action needed:
  • 1 failed job — run `night-watch logs` to investigate
  • 1 PR with failing CI — check PR #43

Implementation

Files Changed

• packages/core/src/utils/summary.ts - New file: getSummaryData() function
• packages/core/src/index.ts - Export from summary.ts
• packages/cli/src/commands/summary.ts - New file: command implementation
• packages/cli/src/cli.ts - Register summary command
• packages/cli/src/__tests__/commands/summary.test.ts - Unit tests

Tests

All 11 tests pass:

• Help text display
• Summary header with time window
• JSON output validation
• Default 12 hours behavior
• Custom --hours value
• No recent activity handling
• Job counts display
• Action item generation for failures
• "No action needed" for healthy jobs
• Action items for failing CI PRs

References

• PRD: docs/PRDs/night-watch/prd-summary-command.md

🤖 Generated with Claude Code

[jonit-dev]

🤖 Implemented by GLM-5

[github-actions]

DiffGuard AI Analysis

AI Review Summary

🏆 Overall Score: 88/100

This PR adds a well-structured summary command for the Night Watch CLI with clean separation between CLI and core logic, comprehensive test coverage, and proper TypeScript typing. Minor improvements around performance and maintainability would elevate it further.

---

✅ Key Strengths

• Clean Architecture: Separates CLI presentation (summary.ts) from data aggregation (core/utils/summary.ts), making both testable and reusable.
• Comprehensive Test Coverage: Tests cover JSON/formatted output, various job states, PR scenarios, edge cases, and help text.
• Good Developer Experience: Clear action items, color-coded output, and both human-readable and JSON output modes.

---

⚠️ Areas for Improvement

• Sequential Async Calls: getSummaryData fetches job analytics, PR info, and queue status sequentially. These independent calls could run in parallel with Promise.all() for faster execution.
• Test Mock Fragility: process.cwd is globally mutated before module import in tests. This works but can cause issues if tests run in parallel or if other tests need the original cwd. Consider using dependency injection or a configurable project dir.
• Duplicate Mocked Calls in Tests: getQueueStatus is mocked with defaults in beforeEach then re-mocked identically in the "should show No action needed" test without changing behavior.

---

📋 Issues Found

Issue Type	Issue Name	Affected Components	Description	Impact/Severity
Performance	Sequential Async Calls	packages/core/src/utils/summary.ts	Three independent async operations (getJobRunsAnalytics, collectPrInfo, getQueueStatus) execute sequentially rather than in parallel.	Low
Maintainability	Global State Mutation	summary.test.ts:31-32	Overwriting process.cwd globally before module import is fragile and may cause test isolation issues.	Low
Maintainability	Redundant Test Mock	summary.test.ts:288-297	getQueueStatus is re-mocked with identical values to beforeEach setup, adding noise without value.	Low

---

🔚 Conclusion

A solid, well-tested feature addition that follows established patterns in the codebase. The parallelization opportunity is minor but worth addressing for CLI responsiveness. Ready to merge with optional polish.

---

Analyzed using z-ai/glm-5

[github-actions]

DiffGuard AI Analysis

AI Review Summary

🏆 Overall Score: 82/100

The PR adds a well-structured "summary" command for a morning briefing feature, with clean separation between CLI and core logic, good TypeScript typing, and comprehensive test coverage. A minor display bug and a few robustness issues prevent a higher score.

---

✅ Key Strengths

• Clean Architecture: Separates CLI presentation logic (summary.ts command) from data aggregation (core/utils/summary.ts), enabling reusability and testability.
• Comprehensive Testing: Tests cover JSON output, formatted output, help text, edge cases (no activity), and action item generation with appropriate mocking.
• Type Safety: Well-defined interfaces (ISummaryData, IJobRunCounts) and proper TypeScript usage throughout.

---

⚠️ Areas for Improvement

• Input Validation: Add validation for the --hours parameter to handle non-numeric strings (produces NaN) and negative values gracefully.
• Cross-Platform Path Handling: getProjectName() uses split('/') which won't work correctly on Windows paths with backslashes—consider using path.basename() instead.
• Test Mock Robustness: The module-level mock of process.cwd is fragile; consider using vi.spyOn(process, 'cwd') within individual tests instead.

---

🐛 Bugs Found

Bug Name	Affected Files	Description	Confidence
Extra Parenthesis in Queue Display	packages/cli/src/commands/summary.ts	Line 163 has an extra closing parenthesis in the template string: for ${projectNames.join(', ')}) should be for ${projectNames.join(', ')}. Output shows an extra ) character.	High 🟢

---

📋 Issues Found

Issue Type	Issue Name	Affected Components	Description	Impact/Severity
Maintainability	Fragile Test Setup	summary.test.ts	Module-level process.cwd replacement before import makes tests dependent on execution order and harder to debug.	Medium
Reliability	Missing Input Validation	summary.ts	parseInt(options.hours || '12', 10) can produce NaN or accept negative numbers without warning.	Low

---

🔚 Conclusion

A solid implementation with good architecture and test coverage. The extra parenthesis bug should be fixed before merge, and the input validation is worth addressing for production robustness.

---

Analyzed using z-ai/glm-5

[jonit-dev]

Night Watch QA Report

Changes Classification

• Type: API (CLI command + core utility)
• Files changed: 5

Test Results

API Tests

• Status: All passing
• Tests: 31 test(s) in 2 file(s)
  • packages/cli/src/__tests__/commands/summary.test.ts — 11 tests
  • packages/core/src/__tests__/utils/summary.test.ts — 20 tests (newly added by QA)

Test Coverage Summary

The QA agent added comprehensive tests for the core getSummaryData utility:

• windowHours: Default and custom time windows
• job counts: Success, failure, timeout, rate_limited, skipped statuses
• action items: Failed jobs, timed out jobs, rate limited jobs, failing CI PRs, pending queue items
• PR integration: Branch patterns, open PRs in response
• queue integration: Pending queue items filtering

Files Added

File	Tests	Description
packages/core/src/__tests__/utils/summary.test.ts	20	Core utility unit tests

---

Night Watch QA Agent

🧪 QA run by GLM-5