Cortex Code CLI bundled skills¶

cortex-code-guide¶

Complete reference guide for Cortex Code CLI: all commands, sessions, MCP integration, keyboard shortcuts, and configuration options.

Ways to use it:

• Learn all available slash commands, keyboard shortcuts, and operational modes

• Configure MCP servers, manage Snowflake connections, and set up integrations

• Understand how to resume sessions, manage context, and use file references with @

developing-with-streamlit¶

Create, edit, debug, and style Streamlit in Snowflake apps: including custom CSS theming, custom components, and packaged component development.

Ways to use it:

• Build a data exploration app with interactive filters, Snowpark queries, and Altair charts

• Apply Snowflake brand CSS theming and custom colors to beautify an existing app

• Create a packaged custom component with HTML/JS for reuse across multiple Streamlit apps

snowflake-notebooks¶

Create and edit Workspace notebooks (.ipynb files) in Snowflake: Snowpark Python cells, SQL cells, data analysis, and debugging.

Ways to use it:

• Create a new notebook for exploratory data analysis combining SQL and Snowpark Python cells

• Upload and deploy an existing Jupyter notebook to the Snowflake Workspace

• Debug a failing notebook cell and fix Snowpark session initialization issues

build-react-app¶

Build React/Next.js applications powered by live Snowflake data: dashboards, analytics tools, and data apps deployed to Snowflake.

Ways to use it:

• Build a customer analytics dashboard with real-time Snowflake data and interactive filters

• Create an internal self-service data explorer app for non-technical business users

• Develop an executive KPI visualization tool with drill-down capabilities

warehouse¶

Manage Snowflake warehouses: Gen2 conversion, credit rates, cost analysis, sizing recommendations, and performance optimization.

Ways to use it:

• Convert existing Gen1 warehouses to Gen2 for improved DML performance and faster resume

• Analyze warehouse credit consumption to identify and right-size over-provisioned warehouses

• Diagnose slow warehouse resume times and tune startup configuration for responsiveness

workload-performance-analysis¶

Analyze SQL query performance via ACCOUNT_USAGE: spilling, partition pruning, cache hit rates, clustering key recommendations, SOS and QAS eligibility.

Ways to use it:

• Identify queries with excessive disk spilling and get concrete fix recommendations

• Analyze partition pruning ratios and suggest better clustering keys for large tables

• Find queries eligible for Search Optimization Service or Query Acceleration Service

deploy-to-spcs¶

Deploy containerized applications to Snowpark Container Services (SPCS): push images to Snowflake registry, create services, configure networking and access.

Ways to use it:

• Package and deploy a Python ML inference service as an SPCS service endpoint

• Push a Docker image to the Snowflake container registry and launch a running service

• Grant role-based access to an SPCS service endpoint for secure consumer access

openflow¶

Openflow (NiFi-based) data integration: deploy and configure connectors, diagnose flow failures, and build custom ingestion pipelines into Snowflake.

Ways to use it:

• Deploy a Google Drive to Snowflake CDC connector using Openflow connector templates

• Diagnose a failing NiFi flow by inspecting processor state, logs, and back-pressure

• Build a custom REST API-to-Snowflake ingestion flow with configurable parameters

dynamic-tables¶

Work with Snowflake Dynamic Tables: create incremental pipelines, optimize target lag, monitor refresh health, and troubleshoot failures.

Ways to use it:

• Convert a scheduled Task pipeline to a Dynamic Table DAG for simpler, declarative maintenance

• Debug a Dynamic Table stuck in UPSTREAM_FAILED state and restore it to healthy

• Set optimal target lag and clustering keys for a high-frequency downstream reporting table

dbt-projects-on-snowflake¶

Manage dbt projects deployed as native Snowflake objects via the snow dbt CLI: not standard dbt workflows. Covers deploy, execute, schedule, and document.

Ways to use it:

• Deploy a dbt project as a Snowflake object and schedule it with a Snowflake Task

• Execute a deployed dbt project and inspect run status via EXECUTE DBT PROJECT

• Generate documentation and lineage for a deployed dbt project from within Snowflake

dcm¶

Database Change Management: create, audit, and debug DCM projects for infrastructure-as-code database schema and role management.

Ways to use it:

• Initialize a new DCM project with DEFINE TABLE, DEFINE SCHEMA, and three-tier role patterns

• Deploy schema changes across dev, staging, and production environments via manifest

• Audit an existing DCM project for drift, misconfigurations, or missing role grants

iceberg¶

Manage Apache Iceberg tables in Snowflake: catalog integrations (Glue, Unity Catalog, Polaris), external volumes, auto-refresh, and write support.

Ways to use it:

• Create a catalog integration with AWS Glue IRC and query external Iceberg tables from Snowflake

• Set up an S3 external volume and create Iceberg tables with ALLOW_WRITES enabled

• Diagnose and fix auto-refresh failures or stale data on Iceberg tables

snowflake-postgres¶

Manage Snowflake Postgres instances: create, suspend, resume, reset credentials, configure network policies, and perform health checks.

Ways to use it:

• Create a new Snowflake Postgres instance and retrieve connection credentials

• Diagnose connectivity issues with an existing instance using health check diagnostics

• Reset credentials and update the network policy allowlist for a Postgres instance

integrations¶

Create and manage all Snowflake integration types: API, catalog, external access, notification, security, and storage integrations.

Ways to use it:

• Create an external access integration to allow UDFs to call external REST APIs

• Set up a notification integration to send pipeline alerts via email or Slack

• Configure a storage integration with S3, Azure Blob, or GCS for external stage access

semantic-view¶

Create and manage semantic views for Cortex Analyst: build views, generate VQR suggestions, add verified queries, metrics, dimensions, and filters.

Ways to use it:

• Create a semantic view for a sales dataset and seed it with verified query representations

• Debug why Cortex Analyst returns incorrect SQL by inspecting the semantic view definition

• Add new calculated metrics and dimension filters to an existing semantic view

dashboard¶

Create and modify interactive Snowflake dashboards with charts, KPI widgets, tables, and markdown: including adding, editing, and fixing widgets.

Ways to use it:

• Build a sales performance dashboard combining multiple charts and KPI summary cards

• Create an executive summary view merging data from multiple Snowflake tables

• Add a new chart widget to an existing dashboard without disturbing other tiles

search-optimization¶

Create and configure Cortex Search Services: build search pipelines from documents on a stage, process them into searchable tables, and deploy Cortex Search.

Ways to use it:

• Build a document Q&A search service over PDFs and text files stored on a Snowflake stage

• Ingest Google Docs, DOCX, audio, or video files into a searchable Cortex Search Service

• Update an existing Cortex Search Service with new documents without rebuilding from scratch

snowpark-python¶

Deploy Snowpark Python workloads to Snowflake: UDFs, UDAFs, UDTFs, and stored procedures using the snow snowpark CLI.

Ways to use it:

• Deploy a Python UDF for custom text processing callable directly from Snowflake SQL

• Create and register a Python stored procedure for complex multi-step ETL logic

• Build and deploy a UDTF (table function) for row-by-row data transformation at scale

snowpark-connect¶

Migrate and validate PySpark workloads to run natively on Snowflake using Snowpark Connect: compatibility analysis and output validation.

Ways to use it:

• Assess a PySpark codebase for Snowpark Connect compatibility and migration effort

• Migrate a PySpark ETL pipeline to run natively on Snowflake compute

• Validate that migrated Spark workloads produce identical outputs to the original PySpark jobs

cortex-agent¶

Create, manage, edit, debug, delete, and chat with Cortex Agents: including multi-tool agents combining Cortex Search and Analyst.

Ways to use it:

• Build a natural language Q&A agent grounded in your Snowflake data warehouse

• Debug a failing agent by inspecting its tool configuration and fixing routing issues

• Deploy a multi-tool agent that combines Cortex Search, Analyst, and custom tools

cortex-ai-functions¶

Use Snowflake Cortex AI Functions for text and document processing: classification, entity extraction, sentiment analysis, translation, embedding, OCR, and document parsing.

Ways to use it:

• Extract structured fields from unstructured invoice PDFs stored on a Snowflake stage

• Classify and sentiment-score customer support tickets at scale using AI_CLASSIFY and AI_SENTIMENT

• Translate multilingual product reviews and summarize them at scale

machine-learning¶

End-to-end ML workflows: model training, Model Registry, feature engineering, time-series forecasting, anomaly detection, and GPU-accelerated jobs.

Ways to use it:

• Train a churn prediction model with Snowpark ML and register it in the Snowflake Model Registry

• Run a GPU-enabled deep learning training job on Snowflake compute via ML Jobs

• Deploy a registered model for batch inference directly within Snowflake SQL

data-quality¶

Monitor data quality using Data Metric Functions (DMFs): table comparisons, dataset popularity tracking, SLA alerting, and quality dashboards.

Ways to use it:

• Set up freshness and row count monitors on critical pipeline tables with SLA alerts

• Compare source and target tables after a migration to validate data integrity

• Detect sudden drops in row counts or spikes in null rates before they reach reports

lineage¶

Analyze data lineage in Snowflake: downstream impact analysis, upstream source tracing, column-level lineage, and root cause debugging.

Ways to use it:

• Trace which downstream tables and dashboards break if a source table schema changes

• Find the upstream origin of a specific column in a complex reporting table

• Investigate unexpected metric values by tracing the full lineage back to the source

error-tables-ops¶

Assess, enable, monitor, and manage Error Tables (DML error logging) across your account: which tables have errors, storage usage, and cleanup.

Ways to use it:

• Identify which tables should have error logging enabled and set it up in bulk

• Analyze DML errors such as NOT NULL violations, string truncation, and constraint failures

• Monitor error table storage and retention, and clean up old error records

data-governance¶

Implement data governance: dynamic masking, row access policies, automated PII classification, GDPR/CCPA compliance tagging, and access auditing.

Ways to use it:

• Auto-classify PII columns across an entire schema using SYSTEM$CLASSIFY

• Apply dynamic masking policies so non-privileged roles see only redacted values

• Audit who accessed sensitive tables in the last 30 days using ACCESS_HISTORY

network-security¶

Recommend, evaluate, and migrate Snowflake network policies: including SaaS-managed rules, hybrid policies, and access history-based recommendations.

Ways to use it:

• Generate network policy recommendations based on 90 days of account access history

• Evaluate a candidate policy before deployment to avoid accidentally locking out users

• Migrate existing custom network policies to use Snowflake-managed SaaS IP rules

trust-center¶

Use Snowflake Trust Center: review security findings, manage scanners, check CIS benchmarks, enable Threat Intelligence, and track remediations.

Ways to use it:

• Run a full security scan and review critical findings against CIS benchmark controls

• Enable Threat Intelligence scanner to detect compromised credentials or anomalous logins

• Remediate a high-severity finding, document the resolution, and track it to closure

key-and-secret-management¶

Manage Tri-Secret Secure, customer-managed keys (CMK/BYOK), key rotation schedules, and periodic data rekeying for compliance.

Ways to use it:

• Activate a customer-managed encryption key (CMK) to enable Tri-Secret Secure

• Rotate an existing CMK and audit the full key change history for compliance records

• Enable periodic data rekeying to meet regulatory encryption refresh requirements

data-cleanrooms¶

Work with Snowflake Data Clean Rooms: set up collaborations, run audience overlap analysis, and activate matched segments without exposing raw PII.

Ways to use it:

• Set up a clean room with a partner to analyze overlapping customer segments securely

• Run audience overlap analysis to find shared customers without sharing raw data

• Activate and export a matched audience segment for a targeted marketing campaign

declarative-sharing¶

Declaratively share data products across Snowflake accounts using application packages with versioning: for Internal Marketplace and cross-account distribution.

Ways to use it:

• Share a curated dataset with another Snowflake account via the Internal Marketplace

• Publish a versioned data product with controlled access tiers for different consumers

• Create an application package for cross-region data distribution with release channels

native-app-consumer¶

Install and configure Snowflake Native Apps as a consumer: grant privileges, approve specifications, configure references, and manage installed apps.

Ways to use it:

• Install a partner Native App from the Snowflake Marketplace into your account

• Grant required database privileges and configure references for a newly installed app

• Approve an app specification and set up data access bindings for the application

native-app-provider¶

Build Snowflake Native Apps: create packages, write manifest files, setup scripts, manage versioning, release channels, and publish to the Marketplace.

Ways to use it:

• Create an app package with a manifest.yml and setup.sql script from scratch

• Add a new app version, register a patch, and push it through a release channel

• Publish a Native App to the Snowflake Marketplace with a consumer-facing listing

cost-intelligence¶

Analyze Snowflake costs end-to-end: credits, budgets, warehouse costs, serverless services, anomaly detection, and cost breakdowns by team or department.

Ways to use it:

• Identify the top 5 most expensive warehouses and right-size them for cost savings

• Set up budget alerts and detect unexpected cost spikes before they escalate

• Generate a chargeback report breaking down Cortex AI, storage, and compute costs by team

organization-management¶

Manage a Snowflake organization: accounts, org users, org-wide spending analysis, security posture, MFA readiness, and ORGANIZATION_USAGE views.

Ways to use it:

• Get a 30-day executive summary of all org accounts including costs, usage, and reliability

• Review the org-wide security posture: MFA adoption, login failures, and auth settings

• Analyze cross-account credit and storage spending trends for a quarterly business review

snowconvert-assessment¶

Analyze SQL Server (or other warehouse) to Snowflake migration projects using SnowConvert assessment reports: waves, exclusions, dynamic SQL.

Ways to use it:

• Parse a SnowConvert assessment report to identify high-risk objects and conversion blockers

• Plan deployment waves based on object complexity scores and inter-object dependencies

• Analyze dynamic SQL patterns that require manual conversion effort and estimate scope

skill-development¶

Create, document, and audit skills for Cortex Code: define triggers, structure multi-step workflows, write sub-skills, and test behavior.

Ways to use it:

• Build a new custom skill from scratch with a SKILL.md, frontmatter, and sub-skill structure

• Audit an existing skill for missing trigger keywords, incomplete workflows, or gaps

• Capture a successful Cortex Code session workflow and convert it into a reusable team skill

Environment setup¶

DAG authoring and testing¶

Data operations¶

Lineage and observability¶