Slash Commands
Install /mx-save, /mx-checkpoint, and /mx-buildcontext as global Claude Code slash commands.
MemNexus provides three slash commands for Claude Code that wrap the most common memory workflows. Instead of remembering MCP tool syntax, you type a slash command and Claude handles the rest.
Install
mx setup skills
This installs three skill files to ~/.claude/skills/mx-*/SKILL.md — global to all projects. Restart Claude Code after installing.
# Preview what would change
mx setup skills --dry-run
# See what's installed
mx setup skills --list
# Force reinstall
mx setup skills --force
# Remove all MemNexus skills
mx setup skills --remove
Skills are version-controlled. When you update the CLI, run mx setup skills again to pick up any skill updates.
Commands
/mx-save — Quick Progress Snapshot
Saves a concise memory using your current conversation context — recent tool calls, files you've been working with, your git branch. Conversation tracking is handled for you: it reuses your current conversation ID or starts a new one.
When to use:
- After completing a task, fixing a bug, or making a decision
- Before switching to a different task
- When you want to record a rationale or trade-off
/mx-checkpoint — Interactive Milestone Checkpoint
Walks through a structured checklist: what was done, decisions made, blockers, and next steps. More thorough than /mx-save — use it at major boundaries.
When to use:
- After completing a multi-day feature
- Before opening a PR or handing off work
- When you want to document trade-offs and alternatives considered
/mx-buildcontext — Session Briefing
Calls build_context with your current working directory and task. Displays active work, key facts, gotchas, and recent activity from your memory store.
When to use:
- At the start of a new session
- After context compaction wipes earlier messages
- When picking up a project you haven't touched recently
How It Works
Slash commands are Claude Code skills — markdown files that instruct Claude how to handle a specific workflow. They call the same MCP tools (create_memory, build_context) that are already available in your session. The difference is UX: you type /mx-save instead of constructing a create_memory() call manually.
All three skills are user-invoked only — Claude will not trigger them automatically.
Example Workflow
Start of session
/mx-buildcontextClaude shows a briefing: what you were working on, key facts, gotchas from past sessions, and recent activity.
After completing a task
/mx-saveClaude generates a memory from your conversation context and saves it with the right content and conversation tracking.
End of feature
/mx-checkpointClaude walks through a structured milestone: what shipped, decisions made, blockers, and next steps.
Requirements
- MemNexus CLI installed and authenticated (
mx auth login) - MemNexus MCP server configured (
mx setup) - Claude Code (slash commands are a Claude Code feature)
Next steps
- Managing Memories — full CRUD operations from the CLI
- CommitContext — automatic memory capture on every commit
- Search — find memories by meaning or keywords