obsidian-memory-architecture

Organize an Obsidian vault as Hermes Agent's durable knowledge layer with routing rules.

58|6|Updated Jul 19, 2026
One-click install
npx skills add https://github.com/AtlasOmnia/hermes-custom-pack --skill obsidian-memory-architecture-atlasomnia
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: obsidian-memory-architecture
Source: https://github.com/AtlasOmnia/hermes-custom-pack/tree/main/skills/obsidian-memory-architecture
Command: npx skills add https://github.com/AtlasOmnia/hermes-custom-pack --skill obsidian-memory-architecture-atlasomnia

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve? AI assistants accumulate facts, conversations, procedures, and documents across many disconnected systems, causing duplication, prompt bloat, and lost knowledge. This Skill defines a clear architecture that routes each type of information to the right layer—Hermes memory, session search, skills, cron output, or an Obsidian vault. ## Core Features & Use Cases - Information Routing: A decision framework that classifies new information into memory providers, session history, skills, vault notes, or transient cache. - Vault Scaffolding: A non-destructive Python script creates a minimal vault structure (Daily, Decisions, Projects, Profiles, Templates) with dry-run and symlink-safety checks. - Multi-Profile Coordination: Ownership maps and read/write boundaries let multiple Hermes profiles share one vault safely. - Use Case: You are migrating an old setup where a giant MEMORY.md acted as the agent's database. Use this Skill to slim built-in memory to bootstrap facts, move procedures into skills, and establish Obsidian as the canonical document store. ## Quick Start Ask the agent to scaffold a new Obsidian vault at your chosen path and set up the routing and ownership map for Hermes memory.

Frequently Asked Questions about obsidian-memory-architecture

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

FAQPage Schema
How do I use Obsidian as long-term memory for an AI agent?

Use Obsidian for canonical documents, decisions, project state, and daily timelines while keeping compact facts in the agent's built-in memory and past conversations in session search. The Skill provides a routing table and a scaffold script that creates the vault structure non-destructively.

How do I scaffold an Obsidian vault for Hermes Agent?

Run scripts/scaffold_vault.py with --vault pointing to your target directory, using --dry-run first to preview changes. The script creates folders like Daily, Decisions, Projects, and Templates without overwriting existing files unless --force is passed.

Does this require a specific Hermes memory provider?

No. The architecture works with built-in Hermes memory alone or with any supported optional memory provider. Check your active layers with hermes memory status rather than assuming a particular backend.

Should I copy chat conversations into Obsidian daily notes?

No. Hermes already stores sessions in a searchable database, so duplicating transcripts creates drift and wastes space. Promote only decisions, canonical state, or human-readable summaries into the vault and use session_search for exact history.

Can multiple Hermes profiles share one Obsidian vault?

Yes. Give each profile an owned directory under Profiles/ and document read/write boundaries in an ownership map at System/Assistant/README.md. The vault is a coordination surface, so operating instructions must still enforce the boundaries.

What should never be stored in the Obsidian vault?

Never store API keys, passwords, tokens, private keys, session cookies, or recovery codes in the vault. Treat synced vaults as data shared with the sync provider, and scan tracked files and Git history for secrets before publishing.