Skip to content
Back to skills

Understand Memory Sync

ASecurity

Sync knowledge graph entities to Memory MCP for cross-session codebase awareness

  • 6 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added September 8, 2026
ai-agentsgonode

Works with

  • mcp

Security analysis

A100/100

Scanned September 8, 2026

npx -y skills add kmshihab7878/claude-code-setup --skill understand-memory-sync --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Understand Memory Sync?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Understand Memory Sync
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/kmshihab7878-understand-memory-sync/badge)](https://www.skillsdirectory.com/skills/kmshihab7878-understand-memory-sync)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: understand-memory-sync
description: Sync knowledge graph entities to Memory MCP for cross-session codebase awareness
argument-hint: [project-path]
---

# /understand-memory-sync

Sync the most important entities from a project's knowledge graph to the Memory MCP server, enabling cross-session codebase recall without re-analysis.

## Instructions

1. **Locate the knowledge graph**:
   - If `$ARGUMENTS` contains a path, use that as the project directory
   - Otherwise, use the current working directory
   - Read `.understand-anything/knowledge-graph.json`
   - If not found, tell the user to run `/understand` first

2. **Extract key entities** (top ~50-100 nodes by architectural importance):
   - All `module` nodes (architectural boundaries)
   - All `class` nodes tagged as `agent`, `core-module`, or `governance`
   - File nodes with >200 lines (complexity hotspots)
   - Nodes referenced in the tour steps
   - The project node itself

3. **Create Memory MCP entities** using `mcp__memory__create_entities`:
   - Entity name: `{project}:{node.name}` (e.g., `CoreMind:CoreMind`)
   - Entity type: map node type → Memory entity type (`module` → `Architecture`, `class` → `Component`, `file` → `File`, `function` → `Function`)
   - Observations: node summary, tags, layer assignment, line count, file path

4. **Create Memory MCP relations** using `mcp__memory__create_relations`:
   - For each `imports` edge between synced nodes → `imports` relation
   - For each `inherits` edge → `extends` relation
   - For each `depends_on` edge between modules → `depends_on` relation
   - For each `contains` edge (module → key class) → `contains` relation

5. **Add a project-level observation** via `mcp__memory__add_observations`:
   - Entity: `CoreMind`
   - Observations:
     - "Knowledge graph: {N} nodes, {M} edges, {L} layers"
     - "Analyzed at: {timestamp}, commit: {hash}"
     - "Dashboard: cd ~/Projects/understand-anything/understand-anything-plugin/packages/dashboard && GRAPH_DIR=/path/to/project npx vite --open"

6. **Report** what was synced:
   ```
   Synced to Memory MCP:
   - {N} entities created
   - {M} relations created
   - Key modules: {list}
   - Key components: {list}

   These entities are now available across sessions via mcp__memory__search_nodes.
   ```

## When to Re-Sync

Re-sync after running `/understand` on new changes (incremental update). The skill should check if entities already exist and update observations rather than creating duplicates.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…