agentsclimarketplace

Vault lint

Skill slogsdon/skills-vault-knowledge/skills/vault-lint

Twenty Obsidian knowledge-work skills for Claude Code: connect, contradict, emerge, trace, ghost, weekly-learnings.

Install
npx -y skills add slogsdon/skills-vault-knowledge --skill vault-lint

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Use when /vault-lint is invoked or when Shane wants a weekly vault health check. Scans Concepts/ and the broader vault for contradictions, orphan pages, missing Concept pages, and stale claims.

SKILL.md

3.4 KB, as published. Nobody here has run it

Skill: /vault-lint

Weekly vault health scan — contradictions, orphans, missing Concept pages, and stale claims — output is a prioritized action list, not a report.

Don't: auto-fix anything — surface for Shane to decide. Don't use this for targeted note exploration — that's the obsidian skill.

Steps

  1. Run the four scans below using obsidian CLI commands via bash
  2. Collate results into a prioritized action list
  3. Present findings grouped by category with specific file names and fix instructions
  4. Do NOT auto-fix anything — surface for Shane to decide

Scan 1 — Contradictions between Concept pages

obsidian search query='Concepts/' limit=50

For each Concept page found, read it. Compare claims across pages. Flag pairs where one page asserts X and another asserts not-X (or a materially different position on the same topic).

Output format:

CONTRADICTION: Concepts/[A] vs Concepts/[B]
  [A] says: "[quote]"
  [B] says: "[quote]"
  → Resolve by: updating whichever is older or merging into a single page

Scan 2 — Orphan pages (no inbound links)

obsidian search query='[[' limit=100

Read each page that contains wiki-links and collect every [[linked page]] reference. Any Concept page that appears in zero link lists is an orphan.

Output format:

ORPHAN: Concepts/[page]
  Created: [date if available]
  → Link from a related note or delete if superseded

Scan 3 — Concepts mentioned 3+ times without a dedicated page

obsidian search query='TERM' limit=20

Run searches for noun phrases that appear frequently across the vault. Count mentions. Any term appearing 3+ times that does NOT have a matching file in Concepts/ is a gap.

Strategy: search for terms surfaced during recent /connect and /trace runs first; also scan Daily Notes for recurring nouns.

Output format:

MISSING CONCEPT: "[term]" (N mentions, no Concepts/[term] page)
  Seen in: [list of note names]
  → Create Concepts/[term] page

Scan 4 — Stale claims superseded by newer sources

obsidian search query='Clippings/' limit=30

For each Concept page, check if any claim is contradicted or significantly updated by a more recent Clipping. Look for:

  • Dates cited in Concept pages that are older than 2 years
  • Clippings that explicitly update or retract a prior claim
  • Concepts referencing statistics or studies where a newer Clipping cites updated data

Output format:

STALE: Concepts/[page]
  Claim: "[quote from concept page]"
  Superseded by: Clippings/[newer source] ([date])
  → Update claim or add caveat in Concepts/[page]

Fallback (if qwen unavailable)

Execute all four scans directly using obsidian search and obsidian read commands via bash. Apply the same logic described above. The scans are bash-executable without Qwen.

Output structure

Present findings in this order, most actionable first:

  1. Contradictions (fix these — they corrupt the knowledge base)
  2. Missing Concepts (create these — reduces future orphan risk)
  3. Orphans (link or delete)
  4. Stale claims (update or caveat)

End with a one-line summary: "N issues found across Concepts/ — X contradictions, Y orphans, Z missing pages, W stale claims."

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.