install
source · Clone the upstream repo
git clone https://github.com/majiayu000/claude-skill-registry-data
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/majiayu000/claude-skill-registry-data "$T" && mkdir -p ~/.claude/skills && cp -r "$T/data/memory-sync" ~/.claude/skills/majiayu000-claude-skill-registry-data-memory-sync && rm -rf "$T"
manifest:
data/memory-sync/SKILL.mdsource content
Memory Sync Skill
Purpose
Maintain continuity across sessions:
- Save current context
- Load previous state
- Update learnings
- Track relationships
Invocation
/memory-sync [action] [options]
Arguments
: save, load, update, or statusaction
: Specific memory file--file [file]
: Include all memory files--full
Examples
/memory-sync save # Save current context /memory-sync load # Load context at session start /memory-sync update learnings # Add to learnings /memory-sync status # Show memory state
Memory Files
context.md
Current operational state:
- What's being worked on
- Recent activity
- Pending decisions
- Notes for next session
learnings.md
Accumulated knowledge:
- What works
- What doesn't
- Patterns discovered
- Technical notes
relationships.md
People and connections:
- Key accounts
- Interaction history
- Potential collaborations
content-log.md
Everything posted:
- Full post history
- Performance data
- Prevents repetition
Workflow
Session Start
/memory-sync load
- Read context.md for current state
- Check what was last worked on
- Note any pending items
- Ready to continue
During Session
- Context updates happen naturally
- Log content when posted
- Note learnings as discovered
- Track relationship interactions
Session End
/memory-sync save
- Update context.md with current state
- Log any new learnings
- Update relationship notes
- Prepare for next session
Context Structure
# Current Context ## Last Updated [DATE] ## Current Session Focus [What we're working on] ## Recent Activity [What was just done] ## What Needs Attention [Priority items] ## Pending Decisions [Open questions] ## Context for Next Session [What to start with] ## Notes [Anything else]
Best Practices
Always Save
- Before ending sessions
- After major activities
- When switching focus
Keep Context Fresh
- Remove stale items
- Update priorities
- Clear completed items
Learn Continuously
- Note what works
- Note what doesn't
- Build knowledge base
Integration
Memory system integrates with:
- All other skills (they read/write context)
- Dashboard (displays memory state)
- Notifications (can alert on stale context)
Data Retention
- context.md: Rolling (recent only)
- learnings.md: Permanent (grows over time)
- relationships.md: Permanent
- content-log.md: Permanent archive