Obsidian Vault Context

v1.0.1

Teach Claude to use an Obsidian vault as a shared workspace with persistent state across sessions. Covers vault navigation, orchestration file management, ou...

0· 134·0 current·0 all-time
byTristin E@tesfandiari1

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for tesfandiari1/obsidian-vault-context.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "Obsidian Vault Context" (tesfandiari1/obsidian-vault-context) from ClawHub.
Skill page: https://clawhub.ai/tesfandiari1/obsidian-vault-context
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install obsidian-vault-context

ClawHub CLI

Package manager switcher

npx clawhub@latest install obsidian-vault-context
Security Scan
VirusTotalVirusTotal
Pending
View report →
OpenClawOpenClaw
Benign
high confidence
Purpose & Capability
The name/description match the content: the SKILL.md and reference docs explain how the agent should navigate, read, and write an Obsidian vault and how to set up optional headless sync. There are no unexpected environment variables, binaries, or installs declared that would be unrelated to the stated purpose.
Instruction Scope
The instructions explicitly direct the agent to read and write orchestration files inside the vault (e.g., _context/status.md, tasks.md, decisions.md) and to append rather than overwrite some files. Boundaries are stated (read/write only inside the vault, treat company/legal as read-only, do not modify skill/agent files). This is coherent, but it does grant the agent authority to create and update files in the user's vault automatically — users should be aware that the agent will persist session state to disk and may update status/tasks during sessions.
Install Mechanism
There is no install spec and no code in the package (instruction-only). The references describe installing Obsidian Headless (npm install -g obsidian-headless) as an optional, user-run step; nothing in the skill installs software automatically.
Credentials
The skill declares no required environment variables, credentials, or config paths. The optional Obsidian Headless reference correctly notes that the user will need Obsidian account credentials and Node.js if they choose headless sync; that requirement is documented in the references and is proportional to enabling bidirectional sync.
Persistence & Privilege
always is false and the skill does not request permanent elevated privileges. The skill instructs the agent to persist state into vault files (its explicit purpose). The only notable privilege is the agent's ability to autonomously write to the vault (the platform default) — this is expected for a vault-syncing skill but worth user attention.
Assessment
This skill is a coherent, instruction-only guide for using an Obsidian vault as the agent's persistent workspace. Before enabling it: (1) ensure the vault does not contain secrets or credentials you don't want the agent to touch; (2) if you enable Obsidian Headless, install it yourself and run it as an unprivileged user, use selective sync and E2EE, and be aware ob login stores credentials locally; (3) back up your vault before allowing automatic writes; (4) if you prefer the agent not to modify files, configure sync as pull-only or disable autonomous writes; and (5) review the orchestration files (status.md, tasks.md, decisions.md) to confirm the agent's update rules match your expectations.

Like a lobster shell, security has layers — review code before you run it.

latestvk971h94jf9mf1dee21rqf7hwvx83nhna
134downloads
0stars
2versions
Updated 1mo ago
v1.0.1
MIT-0

Obsidian Vault Context

The vault is the interface. Not chat. Not a terminal. A shared folder of markdown files that both you and the user can see and edit. Treat it like a shared codebase, not a reference library.

Session Startup

  1. Read _context/context-map.md if it exists — it maps projects to source files.
  2. If no context map, scan top-level folders and infer context from names.
  3. Read _context/status.md to understand what's active, blocked, and completed.
  4. If this is a fresh vault with no orchestration files, create them. See references/orchestration-files.md for templates.

Orchestration Files

Lightweight .md files in _context/ (or _Claude/) that give you persistent state across sessions.

FilePurpose
status.mdWhat's active, blocked, completed. Update every session.
tasks.mdPrioritized task list with #tags. Update when tasks change.
decisions.mdDecision log with context, options, rationale. Append-only.
context-map.mdProject-to-file index (optional).
scratchpad.mdWorking memory, cleared daily (optional).
learnings.mdOperational knowledge across sessions (optional).

Rules: Read before writing. Update immediately, not at session end. Append to decisions and learnings, never overwrite. Keep entries lean — one line per task, one paragraph per decision.

For detailed format and examples, see references/orchestration-files.md.

Vault Navigation

Folder placement carries meaning. Respect the structure when creating files:

PatternMeaning
foundations/ or core/Strategy, vision, frameworks
product/Specs, features, sprints, architecture
company/ or business/Operations, legal, marketing, hiring
research/ or engine/Technical research, analysis
journal/Session logs, daily notes, reviews
decisions/Past decisions with rationale
work/Agent output: drafts, research, deliverables
archive/Superseded files (read-only reference)

Use [[wikilinks]] for all cross-references so the user can click through in Obsidian.

For a recommended starter layout, see references/vault-structure-template.md.

Output Routing

Save work to the vault so the user sees it in Obsidian — don't present everything in chat.

OutputSave To
Research briefswork/research/ or research/
Draftswork/drafts/ or drafts/
Final deliverableswork/output/ or output/
Quick notes_context/scratchpad.md (append)

Add YAML frontmatter (title, type, status, created, updated, tags) to any file that will be referenced again.

Naming: Use descriptive file names (competitor-analysis-march-2026.md), not generic ones (output-1.md). Always update status.md so the user knows what changed.

Bidirectional Collaboration

When Obsidian Headless is configured, changes flow both ways:

  • User -> Agent: User edits in Obsidian (Mac/iPhone/iPad), syncs to server, you read the update.
  • Agent -> User: You write to the vault, syncs to all user devices.

This means neither party needs to explain what the other already wrote down. You write output to the vault. The user reads it in Obsidian. Both can work on the same document asynchronously.

For server sync setup, see references/obsidian-headless-setup.md.

Boundaries

  • Read and write only within the vault directory. Do not access files outside the vault root.
  • Never modify skill files, agent config, or workflow definitions unless the user explicitly asks you to. Your workspace is _context/ and work/ — not the files that define your own behavior.
  • Treat company/legal/ as read-only. Do not create, edit, or delete files in legal directories.
  • Do not store or log credentials. If a workflow requires authentication, defer to the user.

Context Safety

If context compaction triggers mid-session, write working state to disk immediately:

  1. Current progress -> _context/scratchpad.md
  2. State changes -> _context/status.md

These files reconstitute the session after compaction.

Comments

Loading comments...