Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
Added configurable workspace path (--workspace, OPENCLAWORKSPACE, or cwd). Separate organized output into memory/KM subfolder by default. Fixed missing warning() function. Backward compatible. Documentation updated.
What This Skill Does
Parses MEMORY.md and daily memory files from an OpenClaw workspace, classifies each entry by content type (Research, Decision, Insight, Lesson, Pattern, Project, Reference, Tutorial), and writes timestamped markdown files to organized local folders. Content is deduplicated via hash-based state tracking so re-running sync only processes new entries.
Stores knowledge locally without requiring a third-party service like Notion, giving full control over file format and location.
When to use it
- Reviewing all past research notes grouped by topic
- Auditing what lessons were logged over the past week
- Running a nightly cron job to auto-file new memory entries
- Exporting classified entries as JSON for external processing
- Cleaning up orphaned files after bulk memory edits
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Sync my last 7 days of memory entries to organized folders
- 1Reads MEMORY.md and daily files from the past 7 days in the configured workspace
- 2Classifies each entry by content type, domain, certainty, impact, and tags
- 3Checks local-sync-state.json to skip entries already written in a previous sync
- 4Writes new entries as timestamped markdown files into typed subfolders (Research/, Decision/, etc.)
- 5Updates sync state with content hash mappings for deduplication
New markdown files written to ~/.openclaw/workspace/memory/KM/ with per-type index files regenerated