Agent-Skills.md

Agent Skills: Obsidian Search & Discovery

|

UncategorizedID: laurigates/claude-plugins/search-discovery

Repository

laurigates/claude-plugins
laurigatesLicense: MIT
234

Install this agent skill to your local

pnpm dlx add-skill https://github.com/laurigates/claude-plugins/tree/HEAD/obsidian-plugin/skills/search-discovery

Skill Files

Browse the full folder contents for search-discovery.

Download Skill

Loading file tree…

obsidian-plugin/skills/search-discovery/SKILL.md

Skill Metadata

Name
search-discovery
Description
"Obsidian vault search: full-text/grep, tag listing, link traversal, outline, orphan/dead-end detection, broken wikilink audit. Use when exploring backlinks."

Obsidian Search & Discovery

When to Use This Skill

| Use this skill when... | Use the alternative instead when... | |---|---| | Running full-text or grep-style search, tag queries, or backlink traversal against the running vault | Triaging orphaned notes for archival or MOC linking — use vault-orphans | | Listing unresolved wikilinks Obsidian currently flags as broken | Repairing the broken wikilinks once located — use vault-wikilinks | | Inspecting the heading outline of a single note | Reading the full note content — use vault-files | | Querying notes already organised by a Base view | — use bases |

Search, navigate, and audit the vault link graph using the official Obsidian CLI.

Prerequisites

  • Obsidian desktop v1.12.4+ with CLI enabled
  • Obsidian must be running

Search

Path-Only Search

search returns matching file paths only.

# Basic search
obsidian search query="project roadmap"

# Limit and JSON output
obsidian search query="meeting" limit=10 format=json

# Restrict to a folder
obsidian search query="status::active" path=Projects

# Match count only
obsidian search query="TODO" total

# Case-sensitive
obsidian search query="API" case

Grep-Style Context Search

search:context returns path:line: text lines — the agentic-friendly mode.

# Show matching lines with file/line context
obsidian search:context query="TODO"

# JSON-structured context
obsidian search:context query="TODO" format=json

# Restricted to a folder, case-sensitive
obsidian search:context query="API" path=Projects case

# Limit files searched
obsidian search:context query="TODO" limit=20

search:context is the right default for agents — line numbers feed straight into obsidian read + line offsets, or into edit workflows.

Open Search View

# Pop the search panel pre-filled
obsidian search:open query="review needed"

Obsidian Query Operators

The query= value uses Obsidian's standard search syntax — tag:, path:, file:, line:, [property:value], and so on:

obsidian search query="tag:#publish"
obsidian search query="[status:active] tag:#blog"
obsidian search:context query="path:Daily TODO"

Tags

# All tags
obsidian tags

# Tags with counts (default sort=name)
obsidian tags counts

# Sorted by frequency
obsidian tags sort=count counts

# JSON output
obsidian tags format=json

# Tags for the active file (or for file=/path=)
obsidian tags active
obsidian tags file=Recipe

# Tag info (occurrence count + which files)
obsidian tag name=pkm
obsidian tag name=project/active total
obsidian tag name=pkm verbose

Outline (Headings)

# Heading tree of the active file
obsidian outline

# Specific file
obsidian outline file=Recipe

# JSON for navigation
obsidian outline file=Recipe format=json

# Markdown render
obsidian outline file=Recipe format=md

# Just the heading count
obsidian outline file=Recipe total

Links

# Outgoing links from a note (default: active file)
obsidian links file="Architecture Overview"
obsidian links file="Architecture Overview" total

# Backlinks
obsidian backlinks file="API Design"
obsidian backlinks file="API Design" counts
obsidian backlinks file="API Design" format=json

# Unresolved (broken) wikilinks across the vault
obsidian unresolved
obsidian unresolved counts          # with link counts
obsidian unresolved verbose         # include source files
obsidian unresolved format=json

# Notes with no incoming links
obsidian orphans

# Notes with no outgoing links
obsidian deadends

# Counts only
obsidian orphans total
obsidian deadends total

orphans (no incoming) and deadends (no outgoing) target different problems — combine them to find truly disconnected notes.

Common Flags

| Flag | Description | |------|-------------| | format=json | JSON output for machine parsing | | format=tsv, format=csv | Spreadsheet-friendly output | | format=md | Markdown rendering (outline, base:query) | | limit=N | Cap result count | | path=<folder> | Restrict search to a folder | | case | Case-sensitive search | | total | Return only the count | | counts | Include per-result counts | | verbose | Include source files / details | | --copy | Copy result to clipboard |

Agentic Optimizations

| Context | Command | |---------|---------| | Grep-style search (default) | obsidian search:context query="term" | | Path-only search (structured) | obsidian search query="term" format=json | | Match count | obsidian search query="term" total | | Tag frequency analysis | obsidian tags counts sort=count | | Find tagged notes | obsidian tag name=X verbose | | Heading outline (json) | obsidian outline file=X format=json | | Broken link audit | obsidian unresolved counts verbose | | Orphan detection | obsidian orphans | | Dead-end detection | obsidian deadends | | Link graph for note | obsidian links file=X + obsidian backlinks file=X |

Related Skills

  • vault-files — Read content at a path:line returned by search:context
  • properties — Search and filter by frontmatter properties
  • bases — Pre-built queries via Base views
  • bookmarks — Promote a frequent search to a saved bookmark