q

Fast SQLite-based vault search using FTS5 full-text search index

36 stars
Complexity: medium

About this skill

The `/q` skill offers rapid and powerful search capabilities for personal knowledge vaults built upon SQLite with an FTS5 index. Designed to integrate directly into an AI agent's workflow, it allows users to quickly retrieve information from large collections of notes, such as architectural documentation, project notes, or personal knowledge bases. This skill supports a wide array of search patterns, including full-text queries for specific words or phrases, filtering by note type (e.g., ADRs, Projects, Tasks), and searching by specific tags (e.g., `tag:technology/aws`). Additionally, it provides specialized queries for identifying recently modified notes, finding all backlinks to a specific note, or discovering 'orphan' notes that have no inbound links, aiding in knowledge graph maintenance and discovery. By executing optimized SQLite queries, the `/q` skill dramatically reduces search times from seconds to milliseconds, making information retrieval highly efficient. It's an essential tool for anyone managing a structured knowledge base who needs quick, precise access to their stored information directly through their AI agent.

Best use case

This skill is primarily for knowledge workers, developers, architects, or anyone managing a large, structured personal knowledge vault (e.g., a Zettelkasten or project documentation) who needs to quickly find, analyze, and retrieve specific information. It's ideal for scenarios where rapid access to organized data, contextual retrieval of linked information, or identification of knowledge gaps is critical for productivity and decision-making.

Fast SQLite-based vault search using FTS5 full-text search index

Users should expect a markdown-formatted table of relevant note paths, titles, and snippets, retrieved rapidly from their SQLite vault.

Practical example

Example input

/q architecture patterns

Example output

| path                       | match                                                                             |
| :------------------------- | :-------------------------------------------------------------------------------- |
| docs/patterns/design.md    | Design →patterns← are common solutions to recurring software design problems...   |
| docs/architecture/types.md | Explores various architectural →patterns← like microservices, monoliths, etc.    |

When to use this skill

  • When you need to quickly find information in a large SQLite-backed knowledge vault.
  • When performing full-text searches, filtering by note type, or searching by specific tags.
  • When exploring connections between notes (backlinks) or finding unlinked notes (orphans).
  • When you need to retrieve recently modified notes for review or context.

When not to use this skill

  • If you do not have a SQLite-based knowledge vault, specifically one built with `npm run vault:index`.
  • For searching external websites, general web searches, or databases not configured as this vault.
  • If you require complex analytical queries or data transformations beyond simple search and filtering.

Installation

Claude Code / Cursor / Codex

$curl -o ~/.claude/skills/q/SKILL.md --create-dirs "https://raw.githubusercontent.com/DavidROliverBA/ArchitectKB/main/.claude/skills/q/SKILL.md"

Manual Installation

  1. Download SKILL.md from GitHub
  2. Place it in .claude/skills/q/SKILL.md inside your project
  3. Restart your AI agent — it will auto-discover the skill

How q Compares

Feature / AgentqStandard Approach
Platform SupportClaudeLimited / Varies
Context Awareness High Baseline
Installation ComplexitymediumN/A

Frequently Asked Questions

What does this skill do?

Fast SQLite-based vault search using FTS5 full-text search index

Which AI agents support this skill?

This skill is designed for Claude.

How difficult is it to install?

The installation complexity is rated as medium. You can find the installation instructions above.

Where can I find the source code?

You can find the source code on GitHub using the link provided at the top of the page.

Related Guides

SKILL.md Source

# /q - Quick SQLite Search

Fast vault search using the SQLite FTS5 index. Returns results in milliseconds instead of seconds.

## Prerequisites

The SQLite index must be built first:

```bash
npm run vault:index
```

## Usage Patterns

### Full-text Search

```bash
/q architecture patterns
/q "event driven"        # Phrase search
/q kafka integration     # Multiple terms (AND)
```

### Type Filter

```bash
/q type:Adr             # All ADRs
/q type:Project status:active
/q type:Task priority:high
```

### Tag Search

```bash
/q tag:technology/aws
/q tag:project/your-project
```

### Recent Notes

```bash
/q recent               # Modified in last 7 days
/q recent:30            # Modified in last 30 days
```

### Backlinks

```bash
/q backlinks:"Project - Your Project"
/q backlinks:"System - Your System"
```

### Orphans

```bash
/q orphans              # Notes with no backlinks
```

## Implementation

Execute the appropriate SQLite query based on the search pattern.

### Full-text Search Query

```bash
sqlite3 .data/vault.db -markdown "
SELECT n.path, snippet(fts_content,1,'→','←','...',40) as match
FROM fts_content
JOIN notes n ON fts_content.rowid = n.id
WHERE fts_content MATCH '<search_terms>'
ORDER BY rank
LIMIT 20
"
```

### Type Filter Query

```bash
sqlite3 .data/vault.db -markdown "
SELECT path, title, status, priority
FROM notes
WHERE type = '<Type>'
ORDER BY modified DESC
LIMIT 20
"
```

### Tag Search Query

```bash
sqlite3 .data/vault.db -markdown "
SELECT n.path, n.title, n.type
FROM notes n
JOIN tags t ON n.id = t.note_id
WHERE t.tag = '<tag>'
ORDER BY n.modified DESC
LIMIT 20
"
```

### Recent Notes Query

```bash
sqlite3 .data/vault.db -markdown "
SELECT path, title, type, modified
FROM notes
WHERE modified >= date('now', '-<days> days')
ORDER BY modified DESC
LIMIT 30
"
```

### Backlinks Query

```bash
sqlite3 .data/vault.db -markdown "
SELECT n.path, n.title, n.type
FROM notes n
JOIN links l ON n.id = l.source_id
WHERE l.target_path LIKE '%<note_name>%'
ORDER BY n.modified DESC
"
```

### Orphans Query

```bash
sqlite3 .data/vault.db -markdown "
SELECT n.path, n.title, n.type
FROM notes n
LEFT JOIN links l ON n.id = l.target_id
WHERE l.target_id IS NULL
  AND n.type NOT IN ('DailyNote', 'MOC', 'Dashboard', 'Query')
ORDER BY n.modified DESC
"
```

## Performance

| Query Type  | Grep/Glob | SQLite | Improvement |
| ----------- | --------- | ------ | ----------- |
| Full-text   | 5-15 sec  | 0.01s  | ~1000x      |
| Type filter | 3-5 sec   | 0.007s | ~500x       |
| Tag search  | 2-5 sec   | 0.007s | ~500x       |
| Backlinks   | 10+ sec   | 0.01s  | ~1000x      |

## Rebuilding the Index

The index should be rebuilt when vault content changes significantly:

```bash
npm run vault:index      # Full rebuild
npm run vault:stats      # View current statistics
```

## Tips

1. **Combine filters**: `/q type:Adr status:proposed technology/aws`
2. **Use phrases**: `/q "data platform"` for exact matches
3. **Wildcards**: SQLite FTS5 supports `*` wildcards: `/q architect*`
4. **Present results**: Format output as markdown table for readability

Related Skills

dcf

425
from daloopa/investing

Discounted cash flow valuation with sensitivity analysis

Data & ResearchClaude

notebooklm-research

162
from claude-world/notebooklm-skill

Full-autopilot AI research agent powered by Google NotebookLM (notebooklm-py v0.3.4). Ingests sources (URL, text, PDF, DOCX, YouTube, Google Drive), runs deep web research, asks cited questions, and generates 10 native artifact types (audio podcast, video, cinematic video, slide deck, report, quiz, flashcards, mind map, infographic, data table, study guide). Produces original content drafts via Claude, with optional publishing to social platforms via threads-viral-agent integration. Use this skill when the user mentions: NotebookLM, research with sources, create notebook, generate podcast from articles, turn research into content, trending topic research, research pipeline, source-based analysis, cited research answers, generate slides, generate quiz, make flashcards, deep web research, create infographic, compare sources, research report, study guide, source analysis, or knowledge synthesis.

Data & ResearchClaude

Skill: History

154
from ai-analyst-lab/ai-analyst

## Purpose

Data & ResearchClaude

Skill: Explore Data

154
from ai-analyst-lab/ai-analyst

## Purpose

Data & ResearchClaude

RLM (Recursive Language Model) Skill

44
from richardwhiteii/rlm

The RLM (Recursive Language Model) Skill enables AI agents to process extremely large contexts (10M+ tokens) by recursively chunking, processing, and aggregating results, effectively overcoming context window limitations.

Data & ResearchClaude

nblm

26
from magicseek/nblm

This skill allows AI agents, particularly Claude Code, to directly query and manage your Google NotebookLM notebooks, providing source-grounded and citation-backed answers from Gemini.

Data & ResearchClaude

lastXdays

17
from levineam/lastXdays-skill

Researches any given topic across Reddit, X (Twitter), and the broader web within a custom, configurable time window, synthesizing findings and generating expert-level prompts.

Data & ResearchClaude

tavily-search

3891
from openclaw/skills

Use Tavily API for real-time web search and content extraction. Use when: user needs real-time web search results, research, or current information from the web. Requires Tavily API key.

Data & Research

baidu-search

3891
from openclaw/skills

Search the web using Baidu AI Search Engine (BDSE). Use for live information, documentation, or research topics.

Data & Research

notebooklm

3891
from openclaw/skills

Google NotebookLM 非官方 Python API 的 OpenClaw Skill。支持内容生成(播客、视频、幻灯片、测验、思维导图等)、文档管理和研究自动化。当用户需要使用 NotebookLM 生成音频概述、视频、学习材料或管理知识库时触发。

Data & Research

openclaw-search

3891
from openclaw/skills

Intelligent search for agents. Multi-source retrieval with confidence scoring - web, academic, and Tavily in one unified API.

Data & Research

aisa-tavily

3891
from openclaw/skills

AI-optimized web search via AIsa's Tavily API proxy. Returns concise, relevant results for AI agents through AIsa's unified API gateway.

Data & Research