powerbi-duckdb-lakehouse

powerbi-duckdb-lakehouse is a skill for Claude Code, Codex from santoshkanthety/powerbi-agent. It costs 125 tokens per session (1,919 once invoked), scanned A, original, MIT.

A guide to using DuckDB, an embedded SQL engine, to query Parquet and Delta files in a Microsoft Fabric lakehouse or OneLake. It covers local checks and use inside a Fabric notebook.

In plain words
What is it for?
Use it for row counts, distinct-value checks, null rates, date ranges, freshness checks, data-quality checks, and comparing source data with a semantic model.
Why use it?
It can answer many profiling and reconciliation questions without starting a Spark cluster or consuming Fabric capacity. That makes routine data checks easier to run frequently.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one.

Good fit Use it for row counts, distinct-value checks, null rates, date ranges, freshness checks, data-quality checks, and comparing source data with a semantic model.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/santoshkanthety/powerbi-agent/powerbi-duckdb-lakehouse
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add santoshkanthety/powerbi-agent --skill powerbi-duckdb-lakehouse
Clone the repo
git clone --depth 1 https://github.com/santoshkanthety/powerbi-agent

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for powerbi-duckdb-lakehouse

README.md
[![agentmods](https://agentmods.dev/badge/skills/santoshkanthety/powerbi-agent/powerbi-duckdb-lakehouse/github.svg)](https://agentmods.dev/skills/santoshkanthety/powerbi-agent/powerbi-duckdb-lakehouse)
Your own site
<a href="https://agentmods.dev/skills/santoshkanthety/powerbi-agent/powerbi-duckdb-lakehouse"><img src="https://agentmods.dev/badge/skills/santoshkanthety/powerbi-agent/powerbi-duckdb-lakehouse/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for powerbi-duckdb-lakehouse

Your own site · 80×15
<a href="https://agentmods.dev/skills/santoshkanthety/powerbi-agent/powerbi-duckdb-lakehouse"><img src="https://agentmods.dev/badge/skills/santoshkanthety/powerbi-agent/powerbi-duckdb-lakehouse.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 125 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,919 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00125 $0.01919
Opus 5.5 $0.00050 $0.00768
Sonnet 5.5 $0.00025 $0.00384
Haiku 4.5 $0.00013 $0.00192

Measured 16d ago against content hash 846262d0de70, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

powerbi-duckdb-lakehouse scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 16d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/powerbi-duckdb-lakehouse/SKILL.md · 186 lines

How it starts

The opening of the file, as written. The whole thing — 186 lines — stays where its author put it; the contents beside it link to each section on GitHub.

DuckDB over lakehouse and OneLake data

DuckDB is an in-process SQL engine that reads Parquet and Delta directly. Pointed at OneLake, it answers most of the questions people reach for Spark to answer — row counts, distincts, null rates, min/max dates, a reconciliation against what the semantic model reports — with no cluster, no cold start, and no capacity consumption.

That last point is the reason this skill exists. A freshness check that costs nothing can run on every delivery; one that opens a Spark session and smooths CU forward for 24 hours cannot.

Reach for DuckDB first. Reach for powerbi-spark-livy only when the volume or the transformation genuinely needs distributed compute.

Where it runs

Location Use when
Local (your machine, this CLI) Profiling, reconciliation, quality gates, anything in a delivery loop. Needs OneLake access from the host.
Inside a Fabric notebook Data is large enough that pulling it to the host is the bottleneck, but small enough that Spark is overkill. Still cheaper than a Spark session on the same notebook.

Setup

pip install duckdb        # or: uv pip install duckdb

Extensions, loaded per connection:

import duckdb

con = duckdb.connect()
con.execute("INSTALL azure; LOAD azure;")   # ABFSS / OneLake access
con.execute("INSTALL delta; LOAD delta;")   # Delta log awareness — not just the Parquet files

INSTALL needs network access once; on an air-gapped build host, pre-seed the extension directory and skip INSTALL.

Authentication to OneLake

Use the Azure credential chain rather than a key or a SAS token:

con.execute("""
    CREATE SECRET onelake (
        TYPE azure,
        PROVIDER credential_chain,
        CHAIN 'cli;env;managed_identity',
        ACCOUNT_NAME 'onelake'
    );
""")

That picks up az login locally and a managed identity in a hosted runtime. Confirm the exact clause names against the DuckDB azure extension docs for your installed version — the secret syntax has changed across DuckDB releases and a wrong clause fails with a parse error rather than an auth error.

Read the full file on GitHub · 186 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 16d ago First seen · 186 lines · 0 tokens per session scan A 846262d0de70

Subscribe to this mod's changes

powerbi-duckdb-lakehouse is a skill published in the GitHub repository santoshkanthety/powerbi-agent (3 stars, last pushed 17d ago), licensed MIT. It adds 125 tokens to every session and 1,919 once invoked, about $0.0005 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-07.