Obsidian To Expertpack

Convert an existing Obsidian Vault into an agent-ready ExpertPack. Restructures vault content for EK optimization, RAG retrieval, and OpenClaw integration. C...

Security Scan

The skill's code and instructions are consistent with its stated purpose (converting an Obsidian vault into an ExpertPack); it does not request credentials or perform network installs, but it does copy local Obsidian config which can include sensitive personal/plugin data — review output before sharing.

Purpose & Capability

Name/description match the provided artifacts: a Python script plus a migration guide that copy and reformat markdown files into an ExpertPack layout. Required binary is only python3, which is appropriate. No unrelated credentials, binaries, or external services are requested.

Instruction Scope

SKILL.md and references describe reading the vault directory, copying .md files, stripping Dataview blocks, converting links, generating manifests, and copying .obsidian config into the output so the pack opens in Obsidian. All of these actions are within scope for a conversion tool, but copying the .obsidian folder can include user-specific plugin/state data (potentially sensitive). The instructions also ask the user to edit ~/.openclaw/openclaw.json to add the pack to RAG paths — that is expected but modifies a user config file only when manually performed by the user.

Install Mechanism

There is no install spec; this is an instruction-only skill with an included Python script. Nothing is downloaded or executed from external URLs during installation, which minimizes install-time risk.

Credentials

No environment variables, credentials, or unusual config paths are requested by the skill. The script operates on user-supplied filesystem paths only. The only noteworthy item is the intentional copying of the .obsidian directory into the pack output (documented), which is relevant to the goal but may include local plugin settings or other personal data that the user should be aware of.

Persistence & Privilege

The skill is user-invocable and not forced-always. It does not request persistent privileges or modify other skills' configuration. It instructs the user how to add the produced pack path to their OpenClaw config, but that is a user action, not automatic persistence by the skill.

Assessment

This skill appears to do what it says: convert a local Obsidian vault into an ExpertPack using the bundled Python script. Before running: (1) do a --dry-run to preview changes as recommended; (2) review the converted output directory before sharing it — the script copies the .obsidian folder (plugin/state) into the output, which can contain personal/plugin configuration and should be inspected for secrets; (3) back up your source vault first (the tool claims not to modify the source); (4) run the validator tools as instructed to ensure pack integrity; (5) there are no network calls or credential requests in the code, but if you later run additional companion tools (e.g., expertpack-eval) check their requirements. If you want extra caution, open scripts/convert.py and search for any shutil.copytree or write operations to confirm which subfolders are copied and, if desired, edit the script to exclude .obsidian or other sensitive files before conversion.

Obsidian Vault → ExpertPack

Converts an Obsidian Vault into a structured ExpertPack — agent-ready, RAG-optimized, and OpenClaw-compatible. Source vault is never modified; output is a clean copy.

Learn more: expertpack.ai · GitHub

Companion skills: Install expertpack for full EP workflows. Install expertpack-eval to measure EK ratio after conversion.

Step 1: Analyze the Vault

Before running the script, inspect the vault:

1. List the top-level directories — these map to EP content sections
2. Identify the pack type based on structure:
   • journals/, daily/, people/, mind/ → person
   • concepts/, workflows/, troubleshooting/, faq/ → product
   • phases/, checklists/, decisions/, steps/ → process
   • Mix of the above → composite
3. Note any templates/ or _templates/ folders — exclude from conversion
4. Estimate content volume and identify the highest-EK directories

The script auto-detects type (--type auto) but verify your judgment matches before proceeding. See references/migration-guide.md for the full decision tree.

Step 2: Run the Conversion Script

python3 /path/to/ExpertPack/skills/obsidian-to-expertpack/scripts/convert.py \
  /path/to/obsidian-vault \
  --output ~/expertpacks/my-pack-slug \
  --name "My Pack Name" \
  [--type auto|person|product|process|composite] \
  [--dry-run]

Always do a --dry-run first to preview what will be converted.

What the script produces:

• All .md files copied with EP frontmatter (title, type, tags, pack, created)
• Inline #hashtags extracted into frontmatter tags:
• Dataview query blocks stripped (computed views, not knowledge)
• [text](file.md) links converted to [[wikilinks]]
• manifest.yaml, overview.md, glossary.md at pack root
• _index.md in each content directory
• .obsidian/ config copied (pack opens in Obsidian immediately)

For detailed handling of Obsidian-specific patterns (nested tags, daily notes, templates, attachments): read references/migration-guide.md.

Step 3: Validate & Fix

# Fix common issues first
python3 /path/to/ExpertPack/tools/validator/ep-doctor.py ~/expertpacks/my-pack-slug --apply

# Must reach 0 errors
python3 /path/to/ExpertPack/tools/validator/ep-validate.py ~/expertpacks/my-pack-slug --verbose

# Fix any broken wikilinks (cross-vault references)
python3 /path/to/ExpertPack/tools/validator/ep-fix-broken-wikilinks.py ~/expertpacks/my-pack-slug --apply

Do not proceed until ep-validate reports 0 errors.

Step 4: Agent-Assisted Enhancement

After validation, enhance retrieval quality:

1. Lead summaries — add a 1-3 sentence blockquote at the top of the 5-10 most important files
2. Glossary — populate glossary.md with domain-specific terms (this is Tier 1 — always loaded)
3. Propositions — create propositions/ with atomic factual statements extracted from high-EK files
4. EK triage — identify low-EK files (general knowledge) and compress or remove them
5. File size — split files >3KB on ## header boundaries

Step 5: Configure RAG in OpenClaw

Add to ~/.openclaw/openclaw.json:

{
  "agents": {
    "defaults": {
      "memorySearch": {
        "extraPaths": ["/path/to/your/converted-pack"]
      }
    }
  }
}

Restart OpenClaw after config change. The pack is now searchable in every session.

Step 6: Measure EK Ratio

clawhub install expertpack-eval

Run evals to score how much esoteric knowledge the pack contains vs. what the model already knows. Target EK ratio >0.6 for high-value packs.