progress-sync（进度同步）

Progress Sync Expert

I'm the comprehensive progress synchronization expert for SpecWeave. I help you understand and use the /sw:sync-progress command for end-to-end progress synchronization.

⚠️ CRITICAL: AUTO-CREATE IS MANDATORY

When /sw:sync-progress is executed and no external issue exists, it MUST automatically create the issue using the Skill tool.

The command MUST invoke:

• Skill({skill: "sw-github:create", args: "<increment-id>"}) for GitHub
• Skill({skill: "sw-jira:create", args: "<increment-id>"}) for JIRA
• Skill({skill: "sw-ado:create", args: "<increment-id>"}) for Azure DevOps

DO NOT just report "No issues linked" - ACTUALLY CREATE THE ISSUE.

What is /sw:sync-progress?

The TRUE "single button" to sync progress across all systems:

tasks.md → spec.md ACs → living docs → AUTO-CREATE external issues → sync external tools (GitHub/JIRA/ADO)

One command does EVERYTHING - including creating missing external issues!

/sw:sync-progress

No more "No GitHub issue linked" errors! The command auto-creates missing issues.

✅ Archived Increment Behavior

For archived/completed increments, this command ALWAYS creates issues for historical tracking:

Why? Historical tracking is important! Completed work should have external issues for:

• Team visibility
• Sprint retrospectives
• Release notes generation
• Audit trails

For all increments (active or completed): Auto-creates issues if missing (the "single button" philosophy)

When to Use This Command

✅ Use /sw:sync-progress when:

1. First-time sync (no external issue yet): Just created increment, want to sync → auto-creates GitHub/JIRA/ADO issues!
2. After completing tasks: You've marked tasks as done in tasks.md and want to sync everywhere
3. Before closing increment: Final sync before /sw:done to ensure all systems in sync
4. Progress check: Want to update status line and external tools with latest progress
5. After bulk task completion: Completed multiple tasks, sync all at once
6. Manual sync trigger: Hooks didn't fire or you want to force a sync
7. "No GitHub issue linked" error: This command fixes that by auto-creating the issue!

❌ Don't use when:

1. Only want to sync ACs: Use /sw:sync-acs instead (faster, more targeted)
2. Only want to sync docs: Use /sw:sync-specs instead
3. Only want to sync GitHub (issue already exists): Use /sw-github:sync instead
4. Increment not started: No tasks to sync yet
5. Don't want auto-create: Use --no-create flag or manual commands

How It Works

Multi-Phase Orchestration:

Phase 1: Tasks → ACs (spec.md)
  └─ Reads completed tasks from tasks.md
  └─ Finds linked ACs (via "Satisfies ACs" field)
  └─ Marks ACs as complete in spec.md: [ ] → [x]
  └─ Updates metadata.json with AC count

Phase 2: Spec → Living Docs (User Stories)
  └─ Syncs spec.md to living docs structure
  └─ Updates user story completion status
  └─ Generates/updates feature ID if needed

Phase 3: AUTO-CREATE External Issues (NEW!)
  ├─ Checks each configured external tool for linked issues
  ├─ If no issue exists → AUTO-CREATE via /sw-github:create, /sw-jira:create, /sw-ado:create
  ├─ Respects permissions (canUpsertInternalItems, canUpdateExternalItems)
  └─ Skip with --no-create flag if needed

Phase 4: Sync to External Tools (Two-Way)
  ├─ GitHub: Two-way sync (push progress, pull team changes)
  ├─ JIRA: Two-way sync (push tasks, pull status)
  └─ Azure DevOps: Two-way sync (push comments, pull updates)

Phase 5: Status Line Cache
  └─ Updates status line with latest completion %

Usage Examples

Example 1: First-Time Sync (No GitHub Issue Yet) ⭐

Scenario: Just created increment, completed tasks, never created a GitHub issue. Want to sync.

# Single command does EVERYTHING
/sw:sync-progress

What happens:

1. ✅ Tasks → ACs marked complete in spec.md
2. ✅ User stories synced to living docs
3. ✅ GitHub issue AUTO-CREATED (#123)
4. ✅ GitHub issue synced with task progress
5. ✅ Status line shows completion %

No more "No GitHub issue linked" errors!

Example 2: After Completing Tasks (Issue Exists)

Scenario: You completed 5 tasks and marked them in tasks.md. GitHub issue already exists.

# Single command syncs everything
/sw:sync-progress

What happens:

1. ✅ 5 tasks → 12 ACs marked complete in spec.md
2. ✅ 2 user stories marked complete in living docs
3. ✅ GitHub issue #123 detected, synced with progress
4. ✅ Epic issue checklist updated (5/37 tasks complete)
5. ✅ Status line shows 68% → 85% completion

Example 3: Before Closing Increment

Scenario: All 37 tasks complete, ready to close. Ensure final sync.

# Final sync before closure
/sw:sync-progress 0053

# Then close increment
/sw:done 0053

Why important: /sw:done validates completion. Final sync ensures:

• All ACs marked complete
• All user stories synced
• All GitHub issues closed
• Status line shows 100%

Example 4: Dry-Run (Preview Mode)

Scenario: Want to see what will be synced before executing.

# Preview mode
/sw:sync-progress 0053 --dry-run

Output:

🔍 DRY-RUN MODE (No changes made)

Would sync:
   • 37 completed tasks → 70 ACs in spec.md
   • spec.md → 6 user stories in living docs
   • Living docs → 6 GitHub issues (would close completed)
   • Status line cache (would update completion %)

Run without --dry-run to execute sync.

Example 5: Local-Only Sync (No External Tools)

Scenario: Offline work, don't want to sync to GitHub/JIRA yet.

# Skip external tools
/sw:sync-progress 0053 --no-github --no-jira --no-ado

What syncs:

• ✅ Tasks → ACs (spec.md)
• ✅ Spec → Living docs
• ❌ External tools (skipped)
• ✅ Status line cache

Flags

Combine flags:

# Full sync with auto-create (DEFAULT - just works!)
/sw:sync-progress

# Sync only, don't create missing issues
/sw:sync-progress 0053 --no-create

# Dry-run with no external tools
/sw:sync-progress --dry-run --no-github

# Force sync, skip GitHub
/sw:sync-progress --force --no-github

Comparison with Other Sync Commands

Rule of thumb:

• Need complete sync (just works) → Use /sw:sync-progress ✅
• Need targeted sync → Use specific command (sync-acs, sync-specs)
• Need sync only (no auto-create) → Use /sw:sync-progress --no-create

Auto-Detection

Smart increment detection:

# Explicit increment ID
/sw:sync-progress 0053

# Auto-detect from active increment
/sw:sync-progress

How auto-detection works:

1. Reads .specweave/state/active-increment.json
2. Finds first active increment ID
3. Uses that increment for sync

External Tool Configuration

Automatic detection of configured tools:

The command checks .specweave/config.json for:

• GitHub: "provider": "github"
• JIRA: "provider": "jira"
• Azure DevOps: "provider": "azure-devops"

Only configured tools are synced:

✅ GitHub integration detected → Will sync
ℹ️  No JIRA integration → Skip
ℹ️  No ADO integration → Skip

Error Handling

Graceful degradation:

Philosophy: Core sync (tasks → docs) must succeed. External tool sync is best-effort.

Troubleshooting

Issue: "No active increment found"

Error:

❌ No active increment found

Fix:

# Provide increment ID explicitly
/sw:sync-progress 0053

Issue: "AC sync had warnings"

Error:

⚠️  AC sync had warnings: 5 ACs not found in spec.md

Fix:

# Embed ACs from living docs into spec.md
/sw:embed-acs 0053

# Then retry sync
/sw:sync-progress 0053

Why this happens: spec.md missing inline ACs (ADR-0064 requirement).

Issue: "GitHub rate limit exceeded"

Error:

⚠️  GitHub sync had warnings: Rate limit exceeded

Fix: Non-critical. Docs are synced. Retry later when rate limit resets:

# Retry GitHub sync only (when rate limit resets)
/sw-github:sync 0053

Integration with Workflow

Typical increment workflow with progress sync:

# 1. Plan increment
/sw:increment "Safe feature deletion"

# 2. Execute tasks
/sw:do

# [Complete tasks manually or via sub-agents...]

# 3. Sync progress after each batch of tasks
/sw:sync-progress

# 4. Final sync before closure
/sw:sync-progress 0053

# 5. Validate quality
/sw:validate 0053 --quality

# 6. Close increment
/sw:done 0053

Best Practices

✅ DO:

1. Sync after task batches: Complete 3-5 tasks → sync → continue
2. Final sync before closure: Ensure 100% sync before /sw:done
3. Use dry-run first: Preview changes with --dry-run
4. Check external tools: Verify GitHub/JIRA after sync
5. Review status line: Ensure completion % updated correctly

❌ DON'T:

1. Don't sync for every task: Batching is more efficient
2. Don't skip final sync: Always sync before /sw:done
3. Don't ignore warnings: AC sync warnings indicate missing ACs
4. Don't force sync without understanding: --force bypasses validation
5. Don't sync before tasks complete: Sync when progress actually changed

Architecture

Why comprehensive sync is needed:

Problem: Manual multi-step sync is error-prone
  1. Update spec.md ACs manually
  2. Run /sw:sync-specs
  3. Run /sw-github:sync
  4. Run /sw:update-status
  5. Check each system for correctness

Solution: Single command orchestrates all steps
  /sw:sync-progress → Does all 4 steps automatically

Benefits:

• ✅ Single command: One button for complete sync
• ✅ Guaranteed consistency: All systems synced together
• ✅ Error resilience: Non-critical failures don't block core sync
• ✅ Audit trail: Comprehensive report shows what synced
• ✅ Dry-run support: Preview before executing

Background

Before this command, users had to manually:

1. Run /sw:sync-acs
2. Run /sw:sync-specs
3. Run /sw-github:sync
4. Run /sw:update-status

Now: One command does all 4 steps ✅

Related Commands

• /sw:sync-acs - Sync tasks → ACs only
• /sw:sync-specs - Sync spec → living docs only
• /sw:sync-tasks - Sync external → tasks (bidirectional)
• /sw-github:sync - Sync docs → GitHub only
• /sw-jira:sync - Sync docs → JIRA only
• /sw-ado:sync - Sync docs → ADO only
• /sw:update-status - Update status line cache

I'm here to help you sync progress efficiently across all systems!

Ask me:

• "How do I sync progress to GitHub?"
• "What's the difference between sync-progress and sync-acs?"
• "How do I preview sync without executing?"
• "Why did my GitHub sync fail?"
• "When should I use --dry-run?"

Project-Specific Learnings

Before starting work, check for project-specific learnings:

# Check if skill memory exists for this skill
cat .specweave/skill-memories/progress-sync.md 2>/dev/null || echo "No project learnings yet"

Project learnings are automatically captured by the reflection system when corrections or patterns are identified during development. These learnings help you understand project-specific conventions and past decisions.