vault-lint
Skill3 starsChanged 7 months agoApache-2.0View on GitHub
Obsidian vault health checker. Detects broken links, orphan pages, index.md inconsistencies, cross-reference gaps, and frontmatter validation. Inspired by Karpathy's LLM Wiki lint pattern. Triggers: "/vault-lint", "check vault", "vault health", "lint notes"
5 other projects use a version of this.
Adopted by
Most recently taken up by
- jamejialicona-cmyk/vault-startertheir own version, first seen Sep 22, 2026
- Lucioway/way-stacktheir own version, first seen Aug 31, 2026
- slogsdon/skills-vault-knowledgetheir own version, first seen Aug 12, 2026
- tograf/sidebraintheir own version, first seen Jun 21, 2026
- hootbu/llm-wiki-mindtheir own version, first seen May 1, 2026
Other versions5 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 5
- jamejialicona-cmyk/vault-starter★ 2same name
- hootbu/llm-wiki-mind★ 1same name
- tograf/sidebrain★ 1same name
- Lucioway/way-stack★ 0same name
- slogsdon/skills-vault-knowledge★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository5 other files it gives its agents.
- article-to-markdownSkill
- chat-to-obsidianSkill
- github-syncSkill
- review-schedulerSkill
- vault-querySkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
13 rules. 0 are common, 7 are distinctive, and 6 are very specific to this project.
0
Common
Most of the 4 edited versions we read keep these.
7
Distinctive
This author's own choices. Worth a look.
6
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 47
Look for a matching file in $ALL_ARTICLES (filename without .md extension)
- Distinctive: line 48
No match → record as broken link, note the source file and line number
- Very specific: line 54
Search all other articles for [[filename]] or [[filename|
- Distinctive: line 55
No inbound links → record as orphan page
- Very specific: line 65
Check whether total_articles count in the index.md header is accurate
- Distinctive: line 71
Search other articles for these keywords
- Very specific: line 72
If article A frequently mentions core concepts of article B but has no [[B]] link → suggest adding a cross-reference
- Distinctive: line 78
Check entries in index.md:
- Very specific: line 86
Check for YAML frontmatter (starts and ends with ---)
- Distinctive: line 138
Missing index.md entries: read the missing article's frontmatter, append to index.md per the Index Update steps in …
- Very specific: line 139
Count drift in index.md: update total_articles and last_updated
- Distinctive: line 140
Ghost entries: remove from index.md
- Distinctive: line 142
No confirmation needed before auto-fix (these are low-risk metadata corrections).
“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Vault Lint — Knowledge Base Health Check
Inspired by Karpathy's "Lint" philosophy: periodically scan the vault to surface inconsistencies, broken links, orphan pages, and data gaps.
Workflow
Step 1: Environment Detection
1a. Detect vault root — follow ../../_shared/common-steps.md "Vault Detection" section.
1b. Scan all article files:
Glob("**/*.md", path="$VAULT_ROOT")Exclude .claude/, .git/, .playwright-mcp/, reviews/, node_modules/ directories.
Build the file list $ALL_ARTICLES.
Step 2: Six Checks
Check 1: Broken Links
Scan all articles for [[wikilink]] and [[wikilink|alias]]:
Grep("\\[\\[.*?\\]\\]", path="$VAULT_ROOT", glob="**/*.md", output_mode="content")For each wikilink:
- Extract the link target (the part before
|) - Look for a matching file in
$ALL_ARTICLES(filename without.mdextension) - No match → record as broken link, note the source file and line number
Check 2: Orphan Pages
For each file in $ALL_ARTICLES:
- Extract the filename (without path and extension)
- Search all other articles for
[[filename]]or[[filename| - No inbound links → record as orphan page
Exclude index.md, README.md, CLAUDE.md (these naturally have no inbound links).
Check 3: index.md Consistency
- Read
$VAULT_ROOT/index.md - Compare against
$ALL_ARTICLES:- Entry in index.md but file doesn't exist → ghost entry
- File exists but no entry in index.md → missing entry
- Check whether
total_articlescount in the index.md header is accurate
Check 4: Cross-Reference Suggestions
For each article:
- Extract 3-5 core keywords (from title and tags)
- Search other articles for these keywords
- If article A frequently mentions core concepts of article B but has no
[[B]]link → suggest adding a cross-reference
This check is time-intensive; limit to 10 most recently modified articles per run.
Check 5: Summary Quality
Check entries in index.md:
- Summary is empty or too short (< 10 chars) → flag as "insufficient summary"
- Summary doesn't match actual article content (compare against first 50 lines) → flag as "stale summary"
Check 6: Frontmatter Completeness
For each file in $ALL_ARTICLES (excluding index.md, README.md, CLAUDE.md):
- Check for YAML frontmatter (starts and ends with
---) - Validate required fields are present and non-empty:
title,date,tags,type,confidence,related,summary - Validate
typevalue is in the allowed set:concept,deep-dive,comparison,practice,guide,overview - Validate
confidencevalue is in the allowed set:high,medium,low - Validate
relatedis a list with each item in"[[...]]"format (empty list[]is acceptable)
Issue severity:
- Missing frontmatter — file has no YAML frontmatter block
- Missing field — a required field is absent
- Invalid value — type or confidence is not in the allowed enum
Step 3: Generate Health Report
Aggregate results into a structured report, output to terminal:
# Vault Health Report — {date}
## Overview
- Total articles: {count}
- Health status: {healthy / needs attention / needs repair}
## Broken Links ({count})
| Source File | Broken Target | Line |
|------------|---------------|------|
| ... | ... | ... |
## Orphan Pages ({count})
- {filename} — no other article links to this page
## index.md Consistency
- Ghost entries: {list}
- Missing entries: {list}
- Count drift: index says {n}, actual {m}
## Suggested Cross-References
- [[A]] <-> [[B]] (shared keywords: {keywords})
## Summary Quality
- Insufficient: {list}
- Possibly stale: {list}
## Frontmatter Issues ({count})
| File | Issue Type | Details |
|------|-----------|---------|
| ... | ... | ... |Step 4: Auto-Fix (Optional)
Provide auto-fix for these issues:
- Missing index.md entries: read the missing article's frontmatter, append to index.md per the Index Update steps in ../../_shared/common-steps.md
- Count drift in index.md: update
total_articlesandlast_updated - Ghost entries: remove from index.md
No confirmation needed before auto-fix (these are low-risk metadata corrections).
Broken link fixes, orphan page handling, cross-reference additions — report only, do not auto-modify article content.
- Missing or incomplete frontmatter — report only, do not auto-modify (requires content understanding to correctly infer type/confidence).
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

