couchbase-data-modelinglisted
Install: claude install-skill celticht32/Couchbase-Skills-for-Claude.ai
# Couchbase data modeling
A skill for *designing* what to put in Couchbase, not operating an existing cluster. The companion `couchbase-mcp` skill is for executing operations; this one is for the architectural decisions that come before any tool is called.
## When this skill applies
Use this skill whenever the conversation is about *what shape the data should take*, not *what tool to call*. Concrete signals:
- "How should I model X in Couchbase?"
- "Should I embed this or use references?"
- "What's the right key format?"
- "Bucket vs scope vs collection?"
- "I'm coming from [MongoDB / Postgres / DynamoDB] — how do I think about this?"
- "How do I model time-series / events / logs?"
- "Where do I put the embedding vector?"
- "Schema migration in Couchbase?"
If the conversation has already moved to "now run this tool," switch to `couchbase-mcp`. These skills are designed to compose — modeling first, then operation.
## Pick the right reference
| Question | Read |
|---|---|
| "What should my keys look like?" | `references/keys.md` |
| "Should I embed or reference?" / "How big should one document be?" | `references/document-shape.md` |
| "Bucket vs scope vs collection?" | `references/boundaries.md` |
| "How do I model for fast queries / FTS / vector search?" | `references/access-patterns.md` |
| "Time-series, event logs, anything with timestamps and TTL" | `references/time-series-and-ttl.md` |
| "I think I'm doing something wrong" | `references/anti-patterns.md` |
| "I'm co