skills.rest
stanleykao72
stanleykao72COMMUNITY
Explore all skills by this author →
GitHub

memory-obsidian

Store and retrieve agent memories in Obsidian vaults with structured frontmatter.

Updated Mar 1, 2026
Search Tags:#automation#recall#memory#obsidian#vault#knowledge-management#long-term-memory
One-click install
npx skills add https://github.com/stanleykao72/goclaw-plugins --skill memory-obsidian
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: memory-obsidian
Source: https://github.com/stanleykao72/goclaw-plugins/tree/main/skills/memory-obsidian
Command: npx skills add https://github.com/stanleykao72/goclaw-plugins --skill memory-obsidian

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This memory management skill enables long-term memory across multiple agents by storing memories in an Obsidian Vault, allowing per-agent memory spaces and reliable retrieval during and after conversations. It triggers on keywords like 記住、記憶、回憶、忘記、remember、recall、forget、我之前說過、我的偏好、搜尋記憶、記憶統計、列出記憶，並用於對話開始時載入記憶、對話結束前儲存記憶。

Core Features & Use Cases

  • Per-agent memory spaces stored as Markdown files under ~/openclaw-vault/, with unique IDs and structured frontmatter (id, category, importance, scope, tier, etc.).
  • Long-term memory is centralized at {agentId}/long-term/MEMORY.md; daily notes at {agentId}/daily/{YYYY-MM-DD}.md; shared knowledge at _shared/ (read-only).
  • Memory operations via Obsidian CLI: search, read, create, update, delete, with a canonical deduplication step before storing new memories.
  • Memory lifecycle: tiered storage (core, working, peripheral) and a decay/archival process with HEARTBEAT to keep memories relevant.
  • Conversation lifecycle: read long-term at start, optionally search related memories, and append daily summaries and curated memories at end.

Quick Start

Store and recall memories by instructing me what to remember, for example 'Remember Wang主任偏好表格格式'.

Frequently Asked Questions about memory-obsidian

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I store and recall long-term memory for AI agents using an Obsidian vault?▼

Long-term memory is stored as Markdown files with structured frontmatter inside an Obsidian vault. Agents can recall these memories using CLI commands to search and read contextual notes during conversations.

How does deterministic memory creation handle deduplication in knowledge management?▼

Deterministic memory creation applies a canonical deduplication step before storing new memories. This ensures identical or overlapping contextual notes are consolidated rather than duplicated within the vault.

Can I manage per-agent memory spaces with separate directories in a shared vault?▼

Yes, per-agent memory spaces are managed using dedicated directories within a shared vault. Each agent stores long-term notes and daily notes separately, while accessing a read-only shared knowledge directory.

What is the tiered memory lifecycle process for contextual notes?▼

The tiered memory lifecycle categorizes notes into core, working, and peripheral tiers. A decay and archival process with HEARTBEAT manages these tiers to keep stored memories relevant over time.

Do I need Obsidian CLI to automate memory search and update operations?▼

Yes, memory operations such as search, read, create, update, and archive are executed via Obsidian CLI commands. This enables automated retrieval and management of notes during ongoing conversations.

When should I use daily notes versus long-term memory files for agent recall?▼

Daily notes capture summaries of individual conversations, while long-term memory files store curated, persistent knowledge. Long-term memory is loaded at the start of a conversation, with daily notes appended at the end.

Related Skills

Explore More Content & Communication Skills→
Tommy-ywTommy-yw
539

obsidian

Read, search, and create notes in an Obsidian vault.

Community
Basic
jieni777jieni777
1

conversation-memory

Manage persistent short-term, long-term, and entity-based memory for LLM conversations.

Community
Advanced
gancimgancim

para-memory-files

Organizes knowledge into PARA-based memory files with YAML frontmatter for cross-session recall.

Community
Intermediate
woody1234567woody1234567
1

deep-agents-memory

Manages persistent and temporary storage for Deep Agent memory across sessions and threads.

Community
Advanced