Performs CLI semantic search across local files and code repositories.
npx skills add https://github.com/masseater/claude-code-plugin --skill semtools
Please help me install this Agent Skill. Skill: semtools Source: https://github.com/masseater/claude-code-plugin/tree/main/plugins/mutils/skills/semtools Command: npx skills add https://github.com/masseater/claude-code-plugin --skill semtools
semtools provides a CLI for semantic search over local docs and code, enabling you to find relevant items even when exact keywords aren’t present.
Install semtools, then run a search like: search "API design" docs/*.md
High-intent search queries and answers about installing and using this skill.
Semantic search finds relevant content by meaning rather than exact keywords. semtools performs semantic search over local files—PDFs, DOCX, PPTX, Markdown, and source code—using adjustable similarity thresholds to locate matches beyond word-for-word hits, ideal for discovering API patterns or design approaches buried in large codebases.
Yes. semtools searches across PDFs, DOCX, PPTX, Markdown, and source files simultaneously using its search command. The parse command converts PDFs, DOCX, and PPTX to Markdown first via LlamaParse, enabling uniform semantic indexing across mixed document types in your repository.
Install semtools, configure your LlamaParse API key in ~/.semtools_config.json, then run search queries against local files. Use the workspace command to cache results for large-scale searches, and adjust --max-distance and --n-lines flags to control similarity thresholds and context window size.
You need a LlamaParse API key to convert PDFs, DOCX, and PPTX files. Store credentials in ~/.semtools_config.json. For search-only workflows over existing Markdown or source files, no API key is required—only local file access.
Keyword search requires exact term matches; semantic search understands meaning and context. semtools finds conceptually related code—like authentication patterns—even when terminology differs, using adjustable similarity thresholds to balance precision and recall across your codebase.
Yes. The --n-lines flag in semtools adjusts how many surrounding lines appear in results. Use --max-distance to fine-tune the similarity threshold, allowing you to balance between strict matches and broader semantic hits across your documents.
NateBJones-ProjectsConvert heavyweight documents to Markdown or CSV with a lightweight index.
nahisahoSearch HCA Data Portal projects and export CELLxGENE Census subsets for Scanpy analysis.
HokageZParse and extract text, tables, and metadata from PDF, DOCX, HTML, and Markdown files.
ArogyaReddyCombine vector similarity and keyword matching with RRF, Linear, Cross-encoder, or Cascade fusion.
