Skip to content
Get packs

Embedding Strategies

Pick embedding models and chunking methods for semantic search and RAG apps

Source Download skill
Repository
wshobson/agents
Stars
42,095
Licence
MIT
Files
2 files

What it does

When this skill is installed, the agent acts as a guide for choosing and tuning embedding models, the tools that turn text into numerical vectors for search. It can compare models (such as Voyage, OpenAI, and open-source options) by dimensions, token limits, and intended use, and advise on splitting documents into chunks, cleaning text, and avoiding common mistakes like mixing models. It also points to a reference file with fuller templates and worked examples.

Great for

  • Developers building a RAG (retrieval-augmented generation) or semantic search feature who need to choose an embedding model
  • Teams working with domain-specific content such as code, legal, or financial documents
  • Engineers who want guidance on chunk size, overlap, and reducing embedding dimensions

Skip it if

  • People who only need a one-off answer about embeddings and have no application to build; the skill is general guidance and may be dated, since its comparison table is labelled 2026 and pricing or model availability can change
  • Anyone relying on it for exact benchmark quality claims; the file gives recommendations without published evaluation data
  • Production decisions without testing: the advice is generic and must be validated on your own data

Try asking

  • “Which embedding model should I use for a RAG chatbot over our legal contracts?”
  • “How should I chunk long technical manuals before embedding them for search?”
  • “Compare text-embedding-3-small and voyage-3 for a multilingual product catalog.”
  • “Show me a template for embedding documents and storing them with metadata.”
Step-by-step guide: how to use Embedding StrategiesInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 2 original files, the MIT licence and an attribution note.

    embedding-strategies.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip embedding-strategies.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip embedding-strategies.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Give the agent your content type, language, and typical document length so it can narrow the model table to a sensible choice.
  • Ask it to check the reference templates in references/details.md when you need concrete code rather than general advice.
  • Verify current model names, prices, and token limits with the provider before committing, since the table may be out of date.

What's inside

  • references/details.md
  • SKILL.md

Redistributed unmodified from wshobson/agents under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.