# Scripts Directory

Automation scripts for managing Claude configurations and memory.

## sync-memory.sh

Synchronizes Claude's in-memory project knowledge to this version-controlled repository.

### Usage

```bash
# Interactive sync with commit prompt
./scripts/sync-memory.sh

# Auto-commit with timestamp
./scripts/sync-memory.sh --auto

# Preview changes without modifying files
./scripts/sync-memory.sh --dry-run
```

### What It Syncs

- `~/.claude/projects/-source-dlx-src-dlx-ansible/memory/MEMORY.md` → `memory/dlx-ansible/MEMORY.md`
- `~/.claude/projects/-source-hiveops-src-hiveops-mgmt/memory/MEMORY.md` → `memory/hiveops-mgmt/MEMORY.md`
- `~/.claude/projects/-source-hiveops-src-hiveops-incident/memory/MEMORY.md` → `memory/hiveops-incident/MEMORY.md`
- `~/.claude/projects/-source-hiveops-src-hiveops-browser/memory/MEMORY.md` → `memory/hiveops-browser/MEMORY.md`

### Automatic Syncing

#### Option 1: Cron Job (Scheduled)

Run sync every day at 2 AM:

```bash
# Add to crontab
crontab -e

# Add this line:
0 2 * * * cd /source/dlx-src/dlx-claude && ./scripts/sync-memory.sh --auto && git push origin main 2>&1 | logger -t claude-sync
```

#### Option 2: Git Hook (On Commit)

Create a post-commit hook in projects to trigger sync:

```bash
# In ~/.claude/hooks/ (if Claude supports this)
# Or manually run after significant work sessions
```

#### Option 3: Manual (Recommended)

Run manually after completing significant work:

```bash
cd /source/dlx-src/dlx-claude
./scripts/sync-memory.sh
git push origin main
```

### Workflow

1. Work with Claude Code on infrastructure projects
2. Claude updates memory files in `~/.claude/projects/`
3. Run sync script to version control knowledge
4. Push to Gitea for team sharing and backup

### Best Practices

- Run sync after completing major tasks
- Review changes before committing (use interactive mode)
- Use `--dry-run` to preview changes first
- Keep memory files under 200 lines (Claude's limit)
- Document critical learnings immediately

## cleanup-plans.sh

Manages Claude's plan files by archiving old plans to keep the directory clean.

### Usage

```bash
# List all plan files with ages
./scripts/cleanup-plans.sh

# Archive plans older than 30 days
./scripts/cleanup-plans.sh --archive

# Archive all plans
./scripts/cleanup-plans.sh --archive-all

# Change age threshold (e.g., 60 days)
./scripts/cleanup-plans.sh --archive --days 60
```

### What It Does

Claude Code creates plan files in `~/.claude/plans/` for each planning session. Over time, these accumulate. This script:

1. Lists all plan files with their ages
2. Identifies plans older than a threshold (default: 30 days)
3. Archives old plans to `~/.claude/plans-archive/YYYY-MM-DD/`
4. Preserves recent plans in the active directory

### When to Use

- Monthly cleanup: Archive plans older than 30 days
- Project cleanup: Archive all plans when starting fresh
- Regular maintenance: Keep the plans directory manageable

### Archive Location

Archived plans are moved to: `~/.claude/plans-archive/YYYY-MM-DD/`

This preserves plans by archive date, allowing you to:
- Reference old planning sessions if needed
- Restore plans if accidentally archived
- Delete archive folders when truly no longer needed

---

**Created**: 2026-02-09