How does BM25 ranking improve search results?

BM25 is a sophisticated ranking function that estimates document relevance to a query, ensuring that the most contextually significant results appear at the top of your list.

How fast is the search functionality?

Graph-based searches typically complete in approximately 50ms for repositories with over 1,500 notes, providing near-instantaneous feedback.

What happens if a query isn't found in the graph index?

The skill automatically performs a fallback grep search, scanning the full text of all markdown files in your repository to ensure no matches are missed.

Can I filter searches by specific document types?

Yes, you can use filters like 'type:Adr' or shortcuts like 't:Adr' to narrow your search to specific categories of architectural documentation.

Does this skill support regular expressions?

Yes, if a regex pattern is detected in your query, the skill will automatically use grep for a comprehensive content search.

Architect Knowledge Search

Name: Architect Knowledge Search
Author: DavidROliverBA

byDavidROliverBA

•

Aprendizaje y Documentación

Performs intelligent, relevance-ranked searches across architectural knowledge bases using graph indexing and BM25 scoring.

This skill provides a high-performance search interface for the ArchitectKB system, prioritizing structural graph relationships and relevance-based ranking. It efficiently handles metadata filters such as type, status, and priority, along with backlink searches and keyword queries. By querying a pre-built graph index first, it delivers near-instant results for 80% of searches, while automatically falling back to comprehensive grep-based full-text searches to ensure deep content discovery across your entire documentation vault.

Características Principales

01Advanced analysis for orphaned notes and broken internal links

02Automated fallback to full-text grep search for deep content discovery

0310 GitHub stars

04BM25-ranked relevance scoring for highly accurate search results

05Optimized graph queries delivering results in under 50ms

06Multi-parameter filtering by note type, status, and priority

Casos de Uso

01Locating specific Architectural Decision Records (ADRs) or design patterns across large repositories

02Managing knowledge graph health by identifying orphaned notes or stale documentation

03Filtering project documentation and meeting notes by priority or backlink relationships

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add davidroliverba/architectkb search

For use in Claude.ai and ChatGPT

Download Skill