plan-cleanup

name	plan-cleanup
description	Archive completed plans and clean workspace using CLI commands
triggers	["archive plans","clean plans","cleanup workspace","remove old plans","archive completed"]
maintainer	false

Plan Cleanup & Archival Skill

What: Systematically archive completed plans and clean workspace

When to use:

• Plans accumulating in .aiknowsys/ (>20 active plans)
• Completed plans need archival
• Periodic workspace cleanup (monthly/quarterly)
• Before major releases (declutter workspace)
• Plan pointer files getting cluttered

Commands: archive-plans, clean

---

Prerequisites

• AIKnowSys project initialized
• Plan files exist in .aiknowsys/PLAN_*.md
• Plan pointer files in .aiknowsys/plans/active-*.md (optional but recommended)

---

Workflow

Step 1: Assess Current State

Check plan count:

ls -1 .aiknowsys/PLAN_*.md | wc -l

Preview what would be archived:

npx aiknowsys archive-plans --dry-run

Preview full cleanup:

npx aiknowsys clean --dry-run

---

Step 2: Choose Cleanup Strategy

Option A: Archive Completed Plans Only

Use when:

• You want fine-grained control
• Need to archive specific status (COMPLETE, CANCELLED, PAUSED)
• Want to set custom age threshold

Command:

npx aiknowsys archive-plans --status COMPLETE --threshold 0

Options:

• --status <status> - Archive by status (COMPLETE, CANCELLED, PAUSED, etc.) - default: COMPLETE
• --threshold <days> - Archive plans older than N days (use 0 for immediate) - default: 7
• --dry-run - Preview without actually moving files

What it does:

1. Reads plan pointer files (.aiknowsys/plans/active-*.md)
2. Finds plans with matching status and age threshold
3. Moves plan files to .aiknowsys/archived/plans/
4. Updates plan pointer files to remove archived references
5. Shows summary of archived vs kept plans

---

Option B: All-in-One Workspace Cleanup

Use when:

• Monthly/quarterly cleanup routine
• Want to clean everything at once
• Preparing for release or major milestone

Command:

npx aiknowsys clean

What it cleans:

1. ✅ Archives old sessions (>30 days by default)
2. ✅ Archives completed plans (COMPLETE status)
3. ✅ Removes temp files and directories (test artifacts, etc.)

Preview first (recommended):

npx aiknowsys clean --dry-run

---

Step 3: Verify Archival

Check archived plans:

ls -1 .aiknowsys/archived/plans/ | grep PLAN_

Check remaining active plans:

ls -1 .aiknowsys/PLAN_*.md | wc -l

Verify plan pointers updated:

cat .aiknowsys/plans/active-*.md
# Should not reference archived plans

---

Step 4: Commit Changes

Review changes:

git status
# Should show plan files moved to archived/

Commit:

git add -A
git commit -m "chore: archive completed plans (cleanup)

Archived X plans to .aiknowsys/archived/plans/:
- PLAN_feature_x.md (completed)
- PLAN_feature_y.md (completed)
...

Result: N plans → M active plans"

---

Common Scenarios

Scenario 1: Monthly Cleanup Routine

Goal: Declutter workspace at end of month

Steps:

# 1. Preview what would be cleaned
npx aiknowsys clean --dry-run

# 2. Review output, ensure nothing important would be removed

# 3. Run cleanup
npx aiknowsys clean

# 4. Verify results
ls -1 .aiknowsys/PLAN_*.md | wc -l
ls -1 .aiknowsys/archived/plans/ | tail -10

# 5. Commit
git add -A
git commit -m "chore: monthly workspace cleanup"

---

Scenario 2: Archive Specific Status

Goal: Archive all cancelled plans (failed experiments)

Steps:

# 1. Preview
npx aiknowsys archive-plans --status CANCELLED --dry-run

# 2. Archive
npx aiknowsys archive-plans --status CANCELLED --threshold 0

# 3. Verify
ls -1 .aiknowsys/archived/plans/ | grep CANCELLED

# 4. Commit
git add -A
git commit -m "chore: archive cancelled plans"

---

Scenario 3: Pre-Release Cleanup

Goal: Clean workspace before major release

Steps:

# 1. Archive completed plans from this release cycle
npx aiknowsys archive-plans --status COMPLETE --threshold 0

# 2. Archive old sessions (free up space)
npx aiknowsys clean --dry-run  # Preview
npx aiknowsys clean            # Execute

# 3. Verify workspace state
tree .aiknowsys -L 2

# 4. Commit with release context
git add -A
git commit -m "chore: workspace cleanup for v0.11.0 release

Archived completed plans and old sessions
Active plans remaining: $(ls -1 .aiknowsys/PLAN_*.md | wc -l)"

---

Scenario 4: Manual Archival (Without Commands)

When: Commands unavailable or need custom logic

Steps:

# 1. Create archive directory if needed
mkdir -p .aiknowsys/archived/plans

# 2. Move completed plans
mv .aiknowsys/PLAN_feature_x.md .aiknowsys/archived/plans/
mv .aiknowsys/PLAN_feature_y.md .aiknowsys/archived/plans/

# 3. Update plan pointer files manually
# Edit .aiknowsys/plans/active-*.md to remove archived plan references

# 4. Verify and commit
git add -A
git commit -m "chore: archive completed plans (manual)"

⚠️ Limitation: Manual archival doesn't update plan pointer files automatically. Use commands when possible.

---

Best Practices

DO ✅

• Preview first: Always run --dry-run before actual cleanup
• Commit after archival: Keep git history clean
• Archive regularly: Monthly or after major milestones
• Use status-based archival: Archive COMPLETE, CANCELLED separately
• Keep active plans focused: <20 active plans per developer ideal
• Document archival in commit message: List what was archived

DON'T ❌

• Don't archive ACTIVE plans: Use status transitions first (ACTIVE → COMPLETE)
• Don't delete archived plans: Keep for historical reference
• Don't archive without git: Risk losing plan history
• Don't archive plans with uncommitted work: Complete work first
• Don't use aggressive thresholds: --threshold 0 is safe, avoid high values unless certain

---

Integration with Plan Management

Plan Lifecycle:

1. Create plan: npx aiknowsys create-plan --title "Feature X"
2. Work on plan: Update status to ACTIVE
3. Complete work: Update status to COMPLETE
4. Archive plan: npx aiknowsys archive-plans --status COMPLETE
5. Historical reference: Plans remain in .aiknowsys/archived/plans/

Plan Pointer Workflow:

• Plan pointers track active plans per developer
• Archival removes completed plans from pointers automatically
• Pointers keep workspace organized (no manual tracking)

---

Troubleshooting

Archive Command Not Found

Symptom: npx aiknowsys archive-plans → command not recognized

Solution:

# Rebuild CLI
npm run build

# Verify command exists
npx aiknowsys --help | grep archive

---

Plans Not Archived (Kept)

Symptom: archive-plans reports 0 plans archived

Possible causes:

1. Plans don't have COMPLETE status
2. Plans are too recent (within threshold)
3. No plan pointer files exist

Solutions:

# Check plan status
grep -r "status:" .aiknowsys/PLAN_*.md

# Use threshold 0 for immediate archival
npx aiknowsys archive-plans --threshold 0

# Check if plan pointers exist
ls -1 .aiknowsys/plans/active-*.md

---

Accidentally Archived Active Plan

Symptom: Archived a plan still being worked on

Solution:

# Move back from archive
mv .aiknowsys/archived/plans/PLAN_feature_x.md .aiknowsys/

# Update plan pointer if needed
# Edit .aiknowsys/plans/active-<username>.md to re-add plan

# Commit restoration
git add -A
git commit -m "chore: restore accidentally archived plan"

---

Related Commands

• create-plan - Create new implementation plan
• update-plan - Update plan status and content
• query-plans - Query plan metadata
• clean - All-in-one workspace cleanup

---

Related Skills

• context-mutation - Create/update plans and sessions
• context-query - Query plans and sessions

---

See Also

• Plan Management Learned Pattern - Plan workflow patterns
• Multi-Developer Plan Tracking - Plan pointer system

---

Last Updated: 2026-02-09
AIKnowSys Version: 0.10.0+