Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

BigQuery CLI Skill · aiskillstore | DecimalAI

aiskillstore/bigquery

SkillScore95Maintained

Comprehensive guide for using BigQuery CLI (bq) to query and inspect tables in Monzo's BigQuery projects, with emphasis on data sensitivity and INFORMATION_SCHEMA queries.

Databy aiskillstore

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

Data Engineer pack

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/aiskillstore-bigquery/SKILL.md

decimalai skills pull aiskillstore-bigquery --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/aiskillstore-bigquery/SKILL.md https://app.decimal.ai/s/aiskillstore-bigquery/SKILL.md

Raw file: /s/aiskillstore-bigquery/SKILL.md ↗ · machine-readable summary at /s/aiskillstore-bigquery.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

22 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 19d ago

Yes — +55% lift · at higher cost+179% tokens0% turns95% of whole cases pass with skill (21/22)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-13	✗→✓	▲ Improved	199%	0%
case-01	✗→✓	▲ Improved	363%	0%
case-02	✗→✓	▲ Improved	180%	0%
case-03	✗→✓	▲ Improved	121%	0%
case-04	✗→✓	▲ Improved	170%	0%

Showing 5 of 22 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/10/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

aiskillstore / marketplace

skills/bfdcampos/bigquery/SKILL.md

View on GitHub ↗

Imported 8/10/2026Version v1

Skill contents

SKILL.md

BigQuery CLI Skill

This skill provides comprehensive guidance on using the BigQuery CLI (bq) for querying and inspecting data in Monzo's BigQuery projects.

Core Principles

Always specify the project explicitly using --project_id=PROJECT_NAME
Always use Standard SQL with --use_legacy_sql=false
Respect data sensitivity - avoid querying actual content from sensitive tables
Use INFORMATION_SCHEMA for metadata queries (schemas, columns, tables)

Common Query Patterns

1. Check Table Schema (INFORMATION_SCHEMA)

Use this to inspect column names, types, and structure without accessing sensitive data:

bash
bq query --project_id=monzo-analytics --use_legacy_sql=false \
  "SELECT column_name, data_type, is_nullable
   FROM \`monzo-analytics.DATASET_NAME.INFORMATION_SCHEMA.COLUMNS\`
   WHERE table_name = 'TABLE_NAME'
   ORDER BY ordinal_position"

Examples:

bash
# Check dims dataset table schema
bq query --project_id=monzo-analytics --use_legacy_sql=false \
  "SELECT column_name, data_type FROM \`monzo-analytics.dims.INFORMATION_SCHEMA.COLUMNS\`
   WHERE table_name = 'vulnerable_customer_logs_dim' ORDER BY ordinal_position"

# Check prod dataset table schema
bq query --project_id=monzo-analytics --use_legacy_sql=false \
  "SELECT column_name, data_type FROM \`monzo-analytics.prod.INFORMATION_SCHEMA.COLUMNS\`
   WHERE table_name = 'transactions' ORDER BY ordinal_position"

2. Count Rows (Safe for Sensitive Tables)

Use COUNT(*) to check table size without exposing data:

bash
bq query --project_id=monzo-analytics --use_legacy_sql=false \
  "SELECT COUNT(*) as row_count FROM \`monzo-analytics.DATASET.TABLE_NAME\`"

Example:

bash
bq query --project_id=monzo-analytics --use_legacy_sql=false \
  "SELECT COUNT(*) as row_count FROM \`monzo-analytics.dims.vulnerable_customer_logs_dim\`"

3. List All Tables in a Dataset

bash
bq query --project_id=monzo-analytics --use_legacy_sql=false \
  "SELECT table_name, table_type
   FROM \`monzo-analytics.DATASET_NAME.INFORMATION_SCHEMA.TABLES\`
   ORDER BY table_name"

Example:

bash
bq query --project_id=monzo-analytics --use_legacy_sql=false \
  "SELECT table_name FROM \`monzo-analytics.dims.INFORMATION_SCHEMA.TABLES\`
   ORDER BY table_name"

4. Export Schema to File

Useful for programmatic processing of table schemas:

bash
bq query --project_id=monzo-analytics --use_legacy_sql=false \
  --format=csv --quiet \
  "SELECT column_name FROM \`monzo-analytics.DATASET.INFORMATION_SCHEMA.COLUMNS\`
   WHERE table_name = 'TABLE_NAME' ORDER BY ordinal_position" \
  | tail -n +2 > /tmp/columns.txt

Example:

bash
bq query --project_id=monzo-analytics --use_legacy_sql=false \
  --format=csv --quiet \
  "SELECT column_name FROM \`monzo-analytics.dims.INFORMATION_SCHEMA.COLUMNS\`
   WHERE table_name = 'vulnerable_customer_logs_dim' ORDER BY ordinal_position" \
  | tail -n +2 > /tmp/columns.txt

5. Check Table Metadata

Get table creation time, size, and other metadata:

bash
bq query --project_id=monzo-analytics --use_legacy_sql=false \
  "SELECT
     table_name,
     creation_time,
     ROUND(size_bytes/1024/1024/1024, 2) as size_gb,
     row_count
   FROM \`monzo-analytics.DATASET_NAME.INFORMATION_SCHEMA.TABLES\`
   WHERE table_name = 'TABLE_NAME'"

6. Find Tables by Pattern

Search for tables matching a naming pattern:

bash
bq query --project_id=monzo-analytics --use_legacy_sql=false \
  "SELECT table_name
   FROM \`monzo-analytics.DATASET_NAME.INFORMATION_SCHEMA.TABLES\`
   WHERE table_name LIKE '%PATTERN%'
   ORDER BY table_name"

Example:

bash
# Find all customer-related tables
bq query --project_id=monzo-analytics --use_legacy_sql=false \
  "SELECT table_name FROM \`monzo-analytics.dims.INFORMATION_SCHEMA.TABLES\`
   WHERE table_name LIKE '%customer%' ORDER BY table_name"

7. Get Detailed Column Information

Get comprehensive column metadata including descriptions:

bash
bq query --project_id=monzo-analytics --use_legacy_sql=false \
  "SELECT
     column_name,
     data_type,
     is_nullable,
     is_partitioning_column
   FROM \`monzo-analytics.DATASET.INFORMATION_SCHEMA.COLUMNS\`
   WHERE table_name = 'TABLE_NAME'
   ORDER BY ordinal_position"

8. Sample Data (Non-Sensitive Tables Only)

⚠️ WARNING: Only use this on non-sensitive tables. Never query actual content from people/staff/PII tables.

bash
bq query --project_id=monzo-analytics --use_legacy_sql=false \
  "SELECT * FROM \`monzo-analytics.DATASET.TABLE_NAME\` LIMIT 10"

Output Formatting Options

Control how results are displayed:

bash
# CSV format
--format=csv

# JSON format
--format=json

# Pretty table format (default)
--format=prettyjson

# Quiet mode (no status messages)
--quiet

# Maximum rows to return
--max_rows=100

Common Projects and Datasets

Main Analytics Projects

monzo-analytics - Main analytics warehouse
monzo-analytics-v2 - New OOM architecture models
monzo-analytics-pii - PII-containing data (use with caution)
sanitized-events-prod - Sanitised event data
raw-analytics-events-prod - Raw event data

Common Datasets

dims - Dimension tables
prod - Production tables
lending - Lending-specific tables
slurpee - Slurpee data

Data Sensitivity Guidelines

✅ SAFE Operations (Always Allowed)

INFORMATION_SCHEMA queries - These only return metadata, not actual data
COUNT() queries - These only return row counts
Schema inspection - Column names, types, table structure

⚠️ RESTRICTED Operations (Use with Caution)

Querying actual content from:

People/staff data tables
PII-containing tables
Customer financial data
Authentication/security tables

When in doubt:

Stick to INFORMATION_SCHEMA queries
Use COUNT() to verify table exists
Ask the user before querying actual content

🚫 NEVER Do This

Query actual rows from people, staff, hibob tables
Export PII data to local files
Query authentication credentials or tokens
Access customer financial details without explicit permission

Error Handling

Common Errors and Solutions

Error: "Not found: Table"

bash
# Solution: Check the table exists first
bq query --project_id=monzo-analytics --use_legacy_sql=false \
  "SELECT table_name FROM \`monzo-analytics.DATASET.INFORMATION_SCHEMA.TABLES\`
   WHERE table_name LIKE '%SEARCH_TERM%'"

Error: "Access Denied"

bash
# Solution: You may not have permissions for that project/dataset
# Try a different project or ask the user about access

Error: "Syntax error"

bash
# Solution: Ensure you're using Standard SQL (--use_legacy_sql=false)
# Check backtick usage around project.dataset.table identifiers

Best Practices

Always use fully-qualified table names with backticks:

sql project-id.dataset.table

Use LIMIT for exploratory queries to avoid large result sets:

sql SELECT * FROM project.dataset.table LIMIT 10

Check row counts before running expensive queries:

bash # First check size bq query --project_id=monzo-analytics --use_legacy_sql=false \ "SELECT COUNT(*) FROM \project.dataset.table\"

# Then run full query if reasonable

Use dry-run for cost estimation (for expensive queries):

bash bq query --dry_run --use_legacy_sql=false "YOUR_QUERY_HERE"

Export large results to file:

bash bq query --project_id=monzo-analytics --use_legacy_sql=false \ --format=csv "YOUR_QUERY" > output.csv

Quick Reference Commands

bash
# Schema check
bq query --project_id=PROJECT --use_legacy_sql=false \
  "SELECT column_name, data_type FROM \`PROJECT.DATASET.INFORMATION_SCHEMA.COLUMNS\`
   WHERE table_name = 'TABLE' ORDER BY ordinal_position"

# Row count
bq query --project_id=PROJECT --use_legacy_sql=false \
  "SELECT COUNT(*) FROM \`PROJECT.DATASET.TABLE\`"

# List tables
bq query --project_id=PROJECT --use_legacy_sql=false \
  "SELECT table_name FROM \`PROJECT.DATASET.INFORMATION_SCHEMA.TABLES\`
   ORDER BY table_name"

# Table metadata
bq query --project_id=PROJECT --use_legacy_sql=false \
  "SELECT table_name, row_count, size_bytes
   FROM \`PROJECT.DATASET.INFORMATION_SCHEMA.TABLES\`
   WHERE table_name = 'TABLE'"

When to Use This Skill

Invoke this skill when you need to:

Query BigQuery tables or datasets
Inspect table schemas or column types
Count rows or check table existence
Export table metadata
Verify data before running dbt models
Investigate data issues or table structures
Find tables by naming patterns

Integration with dbt Workflow

When working on dbt models in the analytics repository:

Before creating import models - Use BigQuery CLI to inspect source schemas
Before running dbt - Verify source tables exist and have expected structure
Debugging dbt failures - Query actual tables to understand data issues
Validating generators - Check that column types match between source and generator

Remember: Always respect data sensitivity guidelines and use INFORMATION_SCHEMA when possible.

Illustrative Want to see what aiskillstore/bigquery would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Latest benchmark · publisher runVerified

9/3/2026, 20:32 UTC · 22 cases

Lift

+55%

95% vs 41%

Agent turns

1.0 vs 1.0

Tokens

+179%

3,405.23 vs 1,220.68

fail → pass · 12pass → pass · 9fail → fail · 1

Per-case results (22)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-12 I am about to run an exploratory query on `sanitized-events-prod.events.app_launches`. What preliminary step should I run in `bq` CLI before launching the full query?	fail→fail	14,733	11,497	-22%	1	1	0%	1,623	3,704	+128%	0	0	—
case-13 I want to sample 10 rows from a non-sensitive dimension table `monzo-analytics.dims.public_holidays` using `bq`. Write the terminal command.	fail→pass	10,847	7,556	-30%	1	1	0%	1,000	2,991	+199%	0	0	—
case-01 I need to inspect the field names and data types for the `transactions` table in the `prod` dataset using the BigQuery CLI. Could you write the bash command to query this table's structure without pulling any actual row data?	fail→pass	4,654	8,808	+89%	1	1	0%	723	3,351	+363%	0	0	—
case-02 Can you provide a BigQuery CLI command to check the total record count for `vulnerable_customer_logs_dim` in the `dims` dataset? I want a terminal snippet that safely checks size.	fail→pass	7,083	3,463	-51%	1	1	0%	1,168	3,266	+180%	0	0	—
case-03 How can I search for all customer-related tables in the `dims` dataset using the `bq` CLI tool? Please provide a command line query that lists table names matching that pattern.	fail→pass	13,714	8,633	-37%	1	1	0%	1,499	3,309	+121%	0	0	—
case-04 I need to export just the list of column names for `vulnerable_customer_logs_dim` in `dims` to `/tmp/columns.txt` using the BigQuery CLI so another script can ingest it. What bash command line string should I run?	fail→pass	7,659	9,713	+27%	1	1	0%	1,302	3,520	+170%	0	0	—
case-05 I want to retrieve creation time, size in GB, and total row count for the table `transactions` in dataset `prod` using `bq`. Give me the terminal query.	fail→pass	8,129	3,154	-61%	1	1	0%	1,452	3,300	+127%	0	0	—
case-06 Before running a potentially massive query against BigQuery from my terminal, how do I estimate the bytes scanned without actually running or paying for the query execution?	pass→pass	15,934	11,470	-28%	1	1	0%	1,677	3,606	+115%	0	0	—
case-11 When running a `bq query` command, I received a SQL syntax error near backticks in table identifier `monzo-analytics.dims.vulnerable_customer_logs_dim`. What flag is likely missing from my `bq` invocation?	pass→pass	5,669	9,287	+64%	1	1	0%	980	3,308	+238%	0	0	—
case-07 Can you write a `bq query` command that fetches the first 10 rows of user accounts from `monzo-analytics.dims.staff_dim` so I can see what sample data looks like?	fail→pass	5,756	10,529	+83%	1	1	0%	1,031	3,615	+251%	0	0	—
case-08 I want to run a SQL query against BigQuery from my shell to retrieve output in CSV format without any status messages or diagnostic logs printing to stdout. Which flags should I add to `bq query`?	pass→pass	9,179	9,098	-1%	1	1	0%	1,448	3,327	+130%	0	0	—
case-09 How do I get a list of all tables and their table types in the `slurpee` dataset using BigQuery CLI?	fail→pass	9,046	3,520	-61%	1	1	0%	1,474	3,271	+122%	0	0	—
case-10 I need to inspect column names, nullable status, and partitioning column status for table `loans_summary` in dataset `lending` using the BigQuery command line tool. Write the command.	fail→pass	10,051	9,269	-8%	1	1	0%	794	3,338	+320%	0	0	—
case-14 What is the recommended `bq` command pattern to execute a query and export the results directly into a local file named `output.csv`?	pass→pass	13,527	4,348	-68%	1	1	0%	1,569	3,406	+117%	0	0	—
case-15 I am preparing to create a new dbt staging model for source table `raw-analytics-events-prod.raw_events.clickstream`. What BigQuery CLI action should I perform first to verify column names and data types?	fail→pass	12,078	8,927	-26%	1	1	0%	1,221	3,331	+173%	0	0	—
case-16 I need to query the main analytics warehouse project versus the new OOM architecture models project in Monzo's BigQuery setup. Which project IDs should I use for each in the `bq` command?	fail→pass	13,791	8,001	-42%	1	1	0%	1,355	3,173	+134%	0	0	—
case-17 My `bq query` fails with 'Not found: Table monzo-analytics.dims.user_prefs'. How can I construct a CLI command to check if a table with a similar name exists?	fail→pass	14,039	12,708	-9%	1	1	0%	1,581	3,473	+120%	0	0	—
case-18 Which dataset name should be used when querying lending-specific tables in the `monzo-analytics` project using `bq` CLI?	pass→pass	13,228	3,180	-76%	1	1	0%	1,299	3,049	+135%	0	0	—
case-19 When running `bq query` from terminal, what command-line parameter can I pass to cap the maximum number of rows returned in the display output without editing the SQL text?	pass→pass	11,122	3,260	-71%	1	1	0%	967	3,206	+232%	0	0	—
case-20 Write a Python script using `google.cloud.bigquery` library to insert a row into a BigQuery table using streaming inserts.	pass→pass	13,487	9,145	-32%	1	1	0%	1,527	4,411	+189%	0	0	—
case-21 Write a Terraform resource block (`google_bigquery_dataset_iam_binding`) to grant `roles/bigquery.dataViewer` role on dataset `dims` to group `analytics-team@monzo.com`.	pass→pass	3,002	4,428	+48%	1	1	0%	517	3,469	+571%	0	0	—
case-22 Write a dbt model SQL file (`stg_transactions.sql`) using Jinja `source('prod', 'transactions')` function that selects transaction ID, amount, and timestamp.	pass→pass	3,907	5,540	+42%	1	1	0%	648	3,491	+439%	0	0	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +55 percentage points is the difference between those two pass rates over the 22 comparable cases.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 40.9%
Pass rate with the skill: 95.5%
Cases attempted: 22
Cases counted toward lift: 22
Cases that regressed: 0
Skill version measured: v1
Run completed: 9/3/2026, 20:32 UTC

Measured example — same prompt, with vs. without the skill · from the benchmark run · case case-13

Prompt

I want to sample 10 rows from a non-sensitive dimension table `monzo-analytics.dims.public_holidays` using `bq`. Write the terminal command.

Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.

Show both answers

SkillScore blends 4 signals1 of 4 complete

BenchmarkPass rate WITH skill on a verified run95

Live evalNeeds ≥10 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

Efficiency · separate axis (at equal correctness)+0% turns+194% tokens

Shown for transparency; not credited unless a verified run shows no correctness regression.

Backed by 1 of 4 signals — each added signal makes the score more robust.

Verified on gemini-3.6-flash · re-tested 18 days ago

Other measured skills in the registry, with their headline benchmark lift.

Ready to supercharge your AI agents?

aiskillstore/bigquery

SkillSafety

Source

BigQuery CLI Skill

Core Principles

Common Query Patterns

1. Check Table Schema (INFORMATION_SCHEMA)

2. Count Rows (Safe for Sensitive Tables)

3. List All Tables in a Dataset

4. Export Schema to File

5. Check Table Metadata

6. Find Tables by Pattern

7. Get Detailed Column Information

8. Sample Data (Non-Sensitive Tables Only)

Output Formatting Options

Common Projects and Datasets

Main Analytics Projects

Common Datasets

Data Sensitivity Guidelines

✅ SAFE Operations (Always Allowed)

⚠️ RESTRICTED Operations (Use with Caution)

🚫 NEVER Do This

Error Handling

Common Errors and Solutions

Best Practices

Quick Reference Commands

When to Use This Skill

Integration with dbt Workflow

How this skill was measured

Measured example — same prompt, with vs. without the skill · from the benchmark run · case case-13

Related skills