Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
A CLI tool that gives AI agents a typed, file-based memory vault. Memories are stored as Obsidian-compatible markdown using structured templates covering decisions, lessons, procedures, contacts, and more. Checkpoint and recover commands let an agent resume work after a context reset or session death.
Stores memories as plain markdown files that work directly in Obsidian, keeping the vault portable and human-readable without locking data into a proprietary format.
When to use it
- Saving an architectural decision before closing a long session
- Recovering task context after an AI agent restarts unexpectedly
- Logging client contact details and relationship notes across projects
- Searching past decisions by keyword or semantic meaning
- Handing off in-progress work state between agent sessions
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Remember that we decided to use Postgres over SQLite because of concurrent writes
- 1Runs `fsxmemory remember decision "Use Postgres over SQLite"` with the reasoning passed as --content
- 2Memory is written to the decisions/ folder using the structured decision template with YAML frontmatter
- 3Runs `fsxmemory checkpoint --working-on "database selection"` to save current session state
- 4On the next session, runs `fsxmemory recover` to surface the last checkpoint and any handoff notes
- 5Runs `fsxmemory search "database"` to retrieve the stored decision by keyword
Decision stored as a structured markdown file, retrievable by keyword or semantic search in future sessions