Skip to content
Skills Library
AI & agents · Building & coding

vector-store-operations

/vector-store-operations

Run a vector index in production, covering dimensions, filtering, updates, and reindexing when the embedding model changes. Use when semantic search is live and must stay correct as data and models change.

Repository
AI-SKILLS ↗
Licence
MIT
Last repository change
05.09.2026

Assessment and security

90/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content25 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
1
Scripts
0
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 05.09.2026, without references to retired models or APIs.

Licence

MIT, open: use and modify while retaining author attribution.

What it does and when to use it

Run a vector index in production, covering dimensions, filtering, updates, and reindexing when the embedding model changes. Use when semantic search is live and must stay correct as data and models change.

Original author description

Run a vector index in production, covering dimensions, filtering, updates, and reindexing when the embedding model changes. Use when semantic search is live and must stay correct as data and models change.

How it works

  1. Method
  2. Boundaries
Original SKILL.md headings · 38 lines, 248 words
SKILL.md preview
---
name: vector-store-operations
description: Run a vector index in production, covering dimensions, filtering, updates, and reindexing when the embedding model changes. Use when semantic search is live and must stay correct as data and models change.
---

# Vector store operations

A vector index is a derived structure tied to a specific embedding
model. The operational realities that surprise teams are that filters
interact badly with approximate search, updates are not free, and
changing the model means rebuilding everything.

## Method

1. **Pin the embedding model per index.** Vectors from different models
   are not comparable, so a model change is a full reindex rather than a
   rolling update (see embeddings-selection).
2. **Understand your filtering model.** Pre-filtering and post-filtering
   behave very differently with approximate search, and post-filtering
   can return far fewer results than requested.
3. **Store metadata alongside vectors.** Filters on source, date, and
   permissions need to be evaluated in the index rather than after
   retrieval (see realtime-permissions).
4. **Plan updates and deletions explicitly.** Some indexes handle
   deletion by tombstoning and degrade until compacted, which is an
   operational task rather than an automatic one.
5. **Build into a new index and swap.** Reindexing in place leaves the
   system serving inconsistent results during the rebuild (see
   search-indexing-pipeline).
6. **Tune recall against latency deliberately.** Approximate search has
   parameters that trade accuracy for speed, and the defaults are rarely
   right for a specific corpus.
7. **Monitor index size, latency, and recall over time.** All three
   drift as data grows, and recall degradation is silent.

## Boundaries

Vector search finds similar things, which is not always relevant things
(see hybrid-search). Approximate indexes trade guaranteed recall for
speed by design. Vector databases add infrastructure that a modest
corpus may not justify.

Versions

The database has no other skills named vector-store-operations.

Database snapshot: 13.11.2026 · package version 4f339de9c594 (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 100deploy-semantic-search-engine
    Deploy Semantic Search Engine — configure embedding pipeline, vector index, hybrid search, scoring profiles, query expansion, personalization API. Use when: deploy, provision search backend.
  • 98vector-search-patterns
    Implement vector similarity search with embedding generation, index selection, and hybrid retrieval strategies. Covers ChromaDB, pgvector, FAISS, and RAG pipeline design. Triggers on vector search, embeddings, semantic search, or RAG architecture requests.
  • 100rag-engineer
    Expert in building Retrieval-Augmented Generation systems. Masters embedding models, vector databases, chunking strategies, and retrieval optimization for LLM applications. Use when: building RAG, vector search, embeddings, semantic search, document retrieval.
  • 100vector-database
    Use for all vector database work — pgvector schema design, Weaviate collection creation, RAG pipeline scaffolding, embedding model selection, index tuning (HNSW vs IVFFlat), and embedding model migration. Triggers on: 'vector search', 'pgvector', 'weaviate', 'embedding', 'RAG…
  • 100fai-embedding-model-select
    Select and configure embedding models for vector search and RAG pipelines. Use this skill when: - Choosing between embedding models (text-embedding-3-small vs large vs ada-002) - Configuring dimensions, batch sizes, and chunking parameters - Benchmarking embedding quality for a…
  • 100aidp-knowledge-bases
    Build and manage AIDP Knowledge Bases for RAG — create a KB over catalog data, pick an embedding model + chunking, build an HNSW/IVF vector index, add data sources (volume/table), run ingestion jobs, and manage KB permissions. Use when the user wants RAG, a knowledge base,…

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.