semantic-search-enrichment

Enrich regex searches with lemmas, synsets, and optional hypernym chains.

Updated Apr 20, 2026

Search Tags:#semantic search #information retrieval #bm25 #query expansion #wordnet #lemmatization #regex enrichment

Ready to run

npx skills add https://github.com/thistleknot/skills --skill semantic-search-enrichment

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: semantic-search-enrichment
Source: https://github.com/thistleknot/skills/tree/main/semantic-search-enrichment
Command: npx skills add https://github.com/thistleknot/skills --skill semantic-search-enrichment

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Semantic search enrichment fixes missed matches caused by morphological variation and semantic distance, where a literal substring or regex query fails to find conceptually related terms.

Core Features & Use Cases

Linguistic normalization: lemmatizes inflected forms so “configured” matches “configure”.
Synset-based semantic expansion: adds WordNet synonyms to capture near-meaning variants like “bug”→“defect” and “issue”→“problem”.
Hypernym-chain expansion: optionally expands queries to broader concepts (hypernyms) to improve recall during audits and cross-artifact discovery.

Use cases: repository-wide code audits, configuration/wording change detection, and retrieval pipelines where recall matters more than precision.

Quick Start

Use semantic-search-enrichment to expand the term you are searching for, then run the resulting regex pattern against your codebase or document corpus.

Frequently Asked Questions about semantic-search-enrichment

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I improve search recall when literal regex fails to find conceptually related code?▼

Semantic search enrichment improves recall by expanding queries with WordNet lemmas, synonyms, and hypernyms, generating a regex-compatible OR pattern to match morphological variants like configured to configure and synonyms like bug to defect.

How do I match inflected word forms during a repository-wide code audit?▼

To match inflected word forms during a code audit, linguistic normalization lemmatizes inflected words so that searching for configure automatically matches configured, configuring, and configuration variants across the codebase.

What is the best way to find semantically related terms for cross-artifact discovery?▼

The best way to find related terms is using synset-based semantic expansion with optional hypernym chains, adding WordNet synonyms like issue to problem and broader concepts to capture near-meaning variants during discovery.

How do I use semantic search expansion with BM25 scoring pipelines?▼

To use semantic search expansion with BM25 scoring, generate the regex-compatible OR pattern from the expanded lemmas and synonyms, then pass the enriched query downstream to your BM25 scoring or grep pipeline.

What are the limitations of using hypernym chains for query expansion?▼

The limitation of using hypernym chains is that expanding to broader concepts increases false positives, so you must control hypernym depth to limit irrelevant matches and apply caching or batching for performance.