Match task context to relevant tools using embeddings and load them on demand.
npx skills add https://github.com/oimiragieo/LLM-RULES --skill tool-search
Please help me install this Agent Skill. Skill: tool-search Source: https://github.com/oimiragieo/LLM-RULES/tree/main/.claude/skills/tool-search Command: npx skills add https://github.com/oimiragieo/LLM-RULES --skill tool-search
Finding the right tools in large tool libraries is hard and context-heavy. This Skill enables semantic discovery and on-demand loading to dramatically reduce context usage.
Search for a GitHub PR creation tool by task: 'github pull request creation'
High-intent search queries and answers about installing and using this skill.
Semantic tool search uses embeddings to match task context against tool definitions, loading only 3-5 relevant tools on-demand instead of all tools upfront. This achieves 90%+ reduction in tool definition tokens while maintaining access to full tool libraries with 10+ tools.
Yes. The Skill applies semantic matching to tool libraries encompassing GitHub interactions, Jira tickets, and Slack integrations, identifying and loading relevant tools for each task without pre-loading entire integration suites.
Describe your task naturally—for example, 'github pull request creation'—and the embedding-based matching identifies semantically relevant tools from your library, then loads their full definitions on-demand for execution.
The Skill includes graceful error handling to manage cases where matched tools fail to load, while keeping core tools always loaded for reliability so task execution doesn't break on edge cases.
Tool search is optimized for libraries with 10+ tools. At this scale, embedding-based matching and on-demand loading deliver the most significant context savings compared to loading all tool definitions upfront.
team-telnyxGenerate text embeddings and perform semantic search via Telnyx AI APIs.
cosmicdreamsAggregate Jira and Slack activity into a ranked priority triage output.
softmgImplement semantic vector search with HNSW indexing and multiple distance metrics.
tiankong0101-byteRetrieve relevant stored content using hybrid semantic and keyword search.
