memos（memos 搜索）

Memos Skill

⚠️ MANDATORY SKILL INVOCATION ⚠️

YOU MUST invoke this skill (NOT optional) when the user mentions ANY of these triggers:

• "save this to memos", "create a memo", "add a memo"
• "search my memos", "find notes about X", "what did I write about"
• "list my memos", "show recent memos", "what memos do I have"
• "update memo", "edit memo", "delete memo"
• "tag this memo", "add tags", "organize memos"
• "upload to memo", "attach file to memo"
• "archive memo", "pin memo", "make memo public"
• Any mention of "memos", "notes", "note-taking", or "personal knowledge"

Failure to invoke this skill when triggers occur violates your operational requirements.

Purpose

This skill provides read-write access to a self-hosted Memos instance for quick note capture, search, and organization. Memos is a privacy-focused, self-hosted note-taking service with Markdown support, tagging, and file attachments.

Core capabilities:

• Create, read, update, and delete memos (notes)
• Search memos by content, tags, or metadata
• Upload and manage file attachments
• Organize memos with tags
• Archive and visibility controls
• Link related memos together

Primary use case: Quick capture of important information from Claude conversations into a personal knowledge base.

Setup

Prerequisites

• Memos instance running and accessible
• API access token generated from Memos UI
• curl and jq installed

Credential Configuration

Add these variables to ~/.claude-homelab/.env:

# Memos - Self-hosted note-taking service
MEMOS_URL="https://memos.example.com"
MEMOS_API_TOKEN="<your_api_token>"

To generate an API token:

1. Log into your Memos instance
2. Go to Settings → Access Tokens
3. Click "Create" and copy the generated token
4. Add to .env file as shown above

Security:

• .env file is gitignored (never commit)
• Set permissions: chmod 600 ~/.claude-homelab/.env
• Token has same permissions as your user account

Commands

All commands return JSON output for LLM parsing. Scripts source credentials from .env automatically.

Memo Operations

Create a memo:

bash scripts/memo-api.sh create "Your memo content here"
bash scripts/memo-api.sh create "Memo with tags" --tags "work,project"
bash scripts/memo-api.sh create "Private memo" --visibility PRIVATE

List memos:

bash scripts/memo-api.sh list
bash scripts/memo-api.sh list --limit 10
bash scripts/memo-api.sh list --filter 'tag == "work"'

Get specific memo:

bash scripts/memo-api.sh get <memo-id>

Update memo:

bash scripts/memo-api.sh update <memo-id> "Updated content"
bash scripts/memo-api.sh update <memo-id> --add-tags "urgent"

Delete memo:

bash scripts/memo-api.sh delete <memo-id>

Archive memo:

bash scripts/memo-api.sh archive <memo-id>

Search Operations

Search by content:

bash scripts/search-api.sh "search query"
bash scripts/search-api.sh "docker kubernetes" --tags "devops"
bash scripts/search-api.sh "meeting notes" --from "2024-01-01"

Search by tag:

bash scripts/tag-api.sh list                    # List all tags
bash scripts/tag-api.sh search "project-x"      # Find memos with tag

Resource (Attachment) Operations

Upload file:

bash scripts/resource-api.sh upload /path/to/file.pdf
bash scripts/resource-api.sh upload image.png --memo-id <id>

List attachments:

bash scripts/resource-api.sh list
bash scripts/resource-api.sh list --memo-id <id>

Delete attachment:

bash scripts/resource-api.sh delete <attachment-name>

User Operations

Get current user:

bash scripts/user-api.sh whoami

List access tokens:

bash scripts/user-api.sh tokens

Workflow

When the user asks about memos:

1. "Save this to my memos" → Extract key content, create memo with appropriate tags
2. "What did I write about X?" → Search memos by content/tags, present results
3. "Find my notes on project Y" → Use tag search or content filter
4. "Update my memo about Z" → Search for memo, get ID, update content
5. "Delete that memo" → Confirm with user, then delete by ID

Detailed Flow: Quick Capture

User: "Save this conversation about Docker networking to my memos"

1. Extract key information from conversation
2. Create memo with descriptive content
3. Add relevant tags (e.g., "docker", "networking", "conversation")
4. Confirm creation with memo ID
5. Optionally ask if user wants to make it public/private

Detailed Flow: Search and Retrieve

User: "What did I write about Kubernetes last month?"

1. Search memos with query "kubernetes"
2. Apply date filter (last 30 days)
3. Present results with memo IDs and previews
4. User can request full content of specific memos

Detailed Flow: Organization

User: "Tag all my Docker memos with 'devops'"

1. Search for memos containing "docker"
2. For each result, update memo to add "devops" tag
3. Report number of memos updated
4. Show tag statistics

Notes

API Details

• Authentication: Bearer token in Authorization header
• Base URL: /api/v1 endpoint
• Rate limits: No documented limits (self-hosted)
• Pagination: Uses pageSize and pageToken parameters
• Filtering: Google AIP-160 standard (e.g., tag == "work")

Memo Format

Memos support full Markdown syntax:

• Headers, lists, code blocks
• Links and images
• Task lists (- [] and - [x])
• Tables

Visibility Options

• PRIVATE - Only you can see
• PROTECTED - Authenticated users can see
• PUBLIC - Anyone can see (RSS feed)

Best Practices

1. Use descriptive content: First line is preview in UI
2. Tag consistently: Use lowercase, hyphens for multi-word (e.g., "project-alpha")
3. Archive old memos: Keep workspace clean
4. Link related memos: Use memo relations for context

Common Errors

401 Unauthorized:

• Check API token in .env
• Token may have expired (regenerate in Memos UI)

404 Not Found:

• Verify memo ID exists
• Check MEMOS_URL is correct

Connection refused:

• Memos instance not running
• Verify URL in .env

Reference

• Official Docs: https://usememos.com/docs
• API Reference: https://usememos.com/docs/api
• Scripts: skills/memos/scripts/
• Examples: skills/memos/examples/

🔧 Agent Tool Usage Requirements

CRITICAL: When invoking scripts from this skill via the zsh-tool, ALWAYS use pty: true.

Without PTY mode, command output will not be visible even though commands execute successfully.

Correct invocation pattern:

<invoke name="mcp__plugin_zsh-tool_zsh-tool__zsh">
<parameter name="command">./skills/memos/scripts/memo-api.sh [args]</parameter>
<parameter name="pty">true</parameter>
</invoke>

• Troubleshooting: skills/memos/references/troubleshooting.md