Arguments: $ARGUMENTS
Goal: Display task completion status across one or more specs. For 3+ specs, spawn agents in parallel.
Modes
- I001, F001 - Check single infrastructure/feature spec
- A001, W001 - Check single application/website page
- --all - Check all specs
- --infrastructure - Check all infrastructure specs
- --features - Check all feature specs
- --application - Check all Next.js application pages
- --website - Check all Astro marketing/content pages
- (empty) - Check all in-progress specs
Execution
Phase 1: Determine Scope
Goal: Identify which specs to check
Actions:
- Parse $ARGUMENTS to determine mode
- Find all target spec directories:
!{find specs -name "tasks.md" -type f | head -20}
- Count target specs
- Display: "Found N specs to check"
Phase 2: Execute Check
For 1-2 specs: Check directly inline
- For each spec, find tasks.md and count checkboxes:
!{grep -c "^- [ ]" $SPEC_DIR/tasks.md}
!{grep -c "^- [x]" $SPEC_DIR/tasks.md}
- Extract remaining task descriptions:
!{grep "^- [ ]" $SPEC_DIR/tasks.md}
- Display results inline
For 3+ specs: Spawn progress-tracker agents in parallel
- Create one Task() call per spec, send ALL in single message
- Use subagent_type="implementation:progress-tracker"
- Each agent reads tasks.md, counts checkboxes, returns JSON with spec_id, total, completed, remaining, remaining_tasks
- Wait for ALL agents to complete
- Aggregate results
Phase 3: Display Summary
Goal: Show consolidated status
Actions:
- Display per-spec status as table showing: Spec ID, Name, Total, Done, Remaining
- Show overall totals: X tasks, Y done (Z%), W remaining
- For single spec, list all remaining tasks with descriptions
- Suggest next action: /implementation:execute SPEC_ID to continue work