How do I install the Skill Retrieval skill?

Run npx skillmds add majiayu000/skill-retrieval in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Skill Retrieval safe to use?

This skill has not completed SkillMD's automated safety review yet. Capability flags: docs only. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Skill Retrieval?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Skill Retrieval free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Skill Retrieval?

majiayu000 (@majiayu000) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Skill Retrieval by majiayu000 — Agent Skill

Name: Skill Retrieval
Rating: 4.2 (5 reviews)
Author: majiayu000

Skill Retrieval

Semantic search over skill libraries using ColBERT embeddings. Use when: - "find a skill for..." or "search for skills that..." - "what skill handles..." or "which skill can..." - "index my skills" or "rebuild the skill index" - "show me relevant skills for this task"

majiayu000 928a110 2 files · 2.9 KB Updated Sep 9, 2026 567 repo stars

File contents

Skill Retrieval System

ColBERT-based semantic search over skill libraries.

Quick Start

# Build the index (one-time, ~30s per skill with LLM analysis)
voyager skill index --verbose

# Search for skills
voyager skill find "resume where we left off"
voyager skill find "what should I work on next"
voyager skill find "turn this workflow into a skill"

CLIs

`voyager skill index`

Build or update the skill search index.

voyager skill index [OPTIONS]

Options:

--paths PATH — Additional skill directories to index
--output DIR — Index location (default: ~/.skill-index/)
--rebuild — Force rebuild from scratch
--skip-llm — Skip LLM analysis (faster but lower quality)
-v, --verbose — Show progress

`voyager skill find`

Search for relevant skills.

voyager skill find "your query" [OPTIONS]

Options:

-k, --top-k N — Number of results (default: 5)
--index DIR — Index location
--json — Output as JSON

Index Locations

Default: ~/.skill-index/
Override via VOYAGER_SKILL_INDEX_PATH env var
Local dogfooding: --output ./.claude/voyager/skill-index/

Skill Sources

Auto-discovers skills from:

Plugin skills: ./skills/
Local mirror: ./.claude/skills/local/
Generated skills: ./.claude/skills/generated/
User skills: ~/.claude/skills/

Override with CLAUDE_SKILLS_PATH env var.

How It Works

Discovery: Find all SKILL.md files in known locations
Analysis: Use LLM to extract metadata (purpose, triggers, capabilities)
Embedding: Generate searchable text from metadata
Indexing: Build ColBERT index for fast semantic search

Indexing happens once; search is instant (no LLM calls).

See reference.md for implementation details.

---name: skill-retrievaldescription: Semantic search over skill libraries using ColBERT embeddings. Use when: - "find a skill for..." or "search for skills that..." - "what skill handles..." or "which skill can..." - "index my skills" or "rebuild the skill index" - "show me relevant skills for this task"--- # Skill Retrieval System ColBERT-based semantic search over skill libraries. ## Quick Start ```bash # Build the index (one-time, ~30s per skill with LLM analysis) voyager skill index --verbose # Search for skills voyager skill find "resume where we left off" voyager skill find "what should I work on next" voyager skill find "turn this workflow into a skill" ``` ## CLIs ### `voyager skill index` Build or update the skill search index. ```bash voyager skill index [OPTIONS] ``` Options: - `--paths PATH` — Additional skill directories to index - `--output DIR` — Index location (default: `~/.skill-index/`) - `--rebuild` — Force rebuild from scratch - `--skip-llm` — Skip LLM analysis (faster but lower quality) - `-v, --verbose` — Show progress ### `voyager skill find` Search for relevant skills. ```bash voyager skill find "your query" [OPTIONS] ``` Options: - `-k, --top-k N` — Number of results (default: 5) - `--index DIR` — Index location - `--json` — Output as JSON ## Index Locations - Default: `~/.skill-index/` - Override via `VOYAGER_SKILL_INDEX_PATH` env var - Local dogfooding: `--output ./.claude/voyager/skill-index/` ## Skill Sources Auto-discovers skills from: - Plugin skills: `./skills/` - Local mirror: `./.claude/skills/local/` - Generated skills: `./.claude/skills/generated/` - User skills: `~/.claude/skills/` Override with `CLAUDE_SKILLS_PATH` env var. ## How It Works 1. **Discovery**: Find all `SKILL.md` files in known locations 2. **Analysis**: Use LLM to extract metadata (purpose, triggers, capabilities) 3. **Embedding**: Generate searchable text from metadata 4. **Indexing**: Build ColBERT index for fast semantic search Indexing happens once; search is instant (no LLM calls). See `reference.md` for implementation details.

Frequently asked questions