Query
- 809 installs
- 527 repo stars
- Updated April 23, 2026
- duckdb/duckdb-skills
query is a DuckDB Claude Code skill that runs SQL or natural-language questions against attached DuckDB sessions or sandboxed CSV and Parquet files for developers exploring analytics without manual CLI wiring.
About
query is a duckdb-skills module that resolves .duckdb-skills/state.sql session state or falls back to sandboxed ad-hoc mode against :memory: with allowed_paths locked to referenced files. It accepts raw SQL or natural language, generates Friendly SQL using idioms like FROM-first selects, GROUP BY ALL, SELECT * EXCLUDE, PIVOT, ASOF joins, and direct reads from CSV or Parquet globs. Before execution it estimates row counts, warns above 1M rows without LIMIT, and flags tables over 10 GB. A seven-step workflow installs DuckDB if missing, describes schemas with duckdb_tables(), runs duckdb -csv output, handles syntax and extension errors, and suggests LIMIT when results exceed 100 rows. Developers reach for query when probing attached analytics databases or local data files from Claude Code.
- query
- Claude Code
- Enhanced workflow
Query by the numbers
- 809 all-time installs (skills.sh)
- Ranked #481 of 4,347 Backend & APIs skills by installs in the Skillselion catalog
- Data as of Aug 3, 2026 (Skillselion catalog sync)
npx skills add https://github.com/duckdb/duckdb-skills --skill queryAdd your badge
Show developers this skill is listed on Skillselion. Paste this into your README.
| Installs | 809 |
|---|---|
| repo stars | ★ 527 |
| Last updated | April 23, 2026 |
| Repository | duckdb/duckdb-skills ↗ |
How do you query DuckDB databases and local files from an agent?
Support build phase development with Claude Code
Who is it for?
Data engineers and backend developers using duckdb-skills session state or one-off CSV and Parquet analysis inside Claude Code workflows.
Skip if: Teams needing persistent ETL orchestration, non-DuckDB warehouses, or GUI-only BI exploration should use dedicated pipeline or warehouse tools instead.
When should I use this skill?
The user asks to query DuckDB, run SQL on CSV or Parquet files, or explore tables in a duckdb-skills session.
What you get
Executed SQL, CSV result output, schema listings, row-count warnings, and interpreted summaries for natural-language questions.
- CSV query results
- Generated Friendly SQL
- Schema and row-count warnings
By the numbers
- Uses a 7-step query workflow from state resolution through result presentation
- Warns on unbounded queries over 1M rows and 10 GB data sources
Related skills
Lark Openapi ExplorerInstantly explore, test, and generate calls against the full Lark (Feishu) OpenAPI surface without leaving their agent workflow.493k
Lark EventConsume real-time events from Lark/Feishu as structured NDJSON streams inside AI agent workflows.387k16kLark Openapi ExplorerWhen an existing Lark/Feishu skill or CLI command cannot fulfill a specific requirement and they need to discover and invoke the exact native OpenAPI endpoint.386k16k
Just ScrapeQuickly search, crawl, extract structured JSON, or monitor web pages without writing custom scraping code.245k41Lark AppsQuery the current visibility and permission scope of a Lark (Feishu) app without writing HTTP client code.235k16k
Design An InterfaceGenerate multiple radically different interface or API designs in parallel before committing to one.188k198kHow it compares
Use query for agent-driven DuckDB CLI exploration with safety gates instead of hand-running duckdb commands without schema or size checks.
FAQ
What modes does the DuckDB query skill support?
The query skill supports session mode by loading .duckdb-skills/state.sql or a per-project ~/.duckdb-skills state file, and ad-hoc sandboxed mode that queries CSV or Parquet paths through :memory: with allowed_paths restricted to referenced files.
How does the query skill handle very large results?
The query skill probes row counts before execution, warns when sources exceed 1M rows without LIMIT or aggregation, flags tables over 10 GB for runtime risk, and suggests LIMIT when CSV output exceeds 100 rows.