Each example is one scenario: a plain-English brief, the config.yaml it becomes, and the resources that brief needs. Point wxctl at the config to plan, apply, test, and tear it back down. For a hand-authored, step-by-step build, see the worked example. To generate a config from a scenario, see the compose flow. Examples sit on a three-tier ladder: primitives/ (product-neutral building blocks), products/<product>/ (one product, one capability), and solutions/ (spanning products). The Group column shows the ladder position, Industry is a discovery aid, and the Tier badge marks complexity, from Light to Heavy. The calculator-weather-agents starter is covered in the quickstart.
GroupExampleIndustryTierScenarioResources
watsonx-orchestratehr-chatbotCross-industryLightHR chatbot answering policy questions from the handbook and looking up employee recordsagent · knowledge_base · tool
watsonx-orchestrateap-processingFinanceMediumInvoice & AP agent: 3-way matching, duplicate detection, exception routingagent · knowledge_base · tool ×2 · orchestrate_connection
watsonx-orchestrateretail-bank-supportFinanceMediumRetail-bank support agent answering product/fee questions and looking up accounts on a custom gateway modelagent · model · knowledge_base · tool ×2 · orchestrate_connection
watsonx-orchestrateclaims-intake-flowInsuranceMediumClaims-intake assistant validating eligibility with a python tool, an OpenAPI tool, and a decisions flow on a gateway modelagent · model · tool (python) · tool (openapi) · tool (flow) · space · orchestrate_connection
watsonx-orchestratemodel-routerCross-industryMediumMulti-model gateway routing: a fast triage agent and a deep specialist behind a supervisor that routes by complexityagent ×3 · model ×2 · space · orchestrate_connection
watsonx-orchestrateops-mcp-assistantCross-industryHeavyOps assistant backed by a custom local MCP server that wxctl packages and uploadsagent · model · toolkit (local MCP) · space · orchestrate_connection
watsonx-aicredit-risk-modelFinanceMediumDeploy a transparent credit-risk scoring model on watsonx.ai and score a loan applicantspace · package_extension · software_specification · wml_function · wml_script · ai_service · wml_deployment
watsonx-datalakehouse-analyticsRetailHeavySaaS retail lakehouse: land CSVs in COS, register Iceberg + Db2 catalogs, query with Presto and Spark, run a Spark ingestion jobstorage_connection · s3_bucket · s3_object · catalog · storage_registration · database_connection · database_registration · presto_engine · spark_engine · schema · ingestion_job
watsonx-datalakehouse-enginesRetailHeavySoftware (CP4D) retail lakehouse with the full engine zoo: Presto, Prestissimo, external Db2, a generic engine, and a Spark ingestion jobstorage_connection · s3_bucket · storage_registration · database_connection · database_registration · presto_engine · prestissimo_engine · db2_engine · other_engine · ingestion_job
watsonx-datasal-enrichmentFinanceHeavyEnable the watsonx.data Semantic Automation Layer and auto-enrich a customer table against a glossary (Software / CP4D only)integration · sal_integration · sal_glossary · sal_global_settings · sal_enrichment_settings · sal_enrichment_job
knowledge-catalogdata-glossaryFinanceHeavyCustomer-data governance taxonomy: category hierarchy, business term, bulk term import, and data-protection rulescategory ×2 · business_term · business_terms · rule · rules
solutionscredit-risk-governanceFinanceHeavyFour products in one config: deploy a credit-risk model, govern it with an OpenScale quality monitor, and front it with an Orchestrate loan-decision agentspace · software_specification · wml_function · wml_deployment · service_provider · data_mart · subscription · monitor_instance · orchestrate_connection ×2 · model · tool · agent

​
Run an example

Every example is self-contained and credential-free: secrets are ${env:VAR} placeholders, never literals. Configure a profile, export any env vars the example lists, then from the example directory: 
wxctl plan    -f config.yaml   # preview the DAG, no credentials needed
wxctl apply   -f config.yaml   # create the resources
wxctl test    -f config.yaml   # run the kind: test checks
wxctl destroy -f config.yaml   # tear it all down
See the examples directory on GitHub for the full briefs, source, and a template for adding your own.

​
Next steps

Worked example

Build a tool, an agent, and a test by hand, then plan, apply, and test.

Automation & CI

Run these configs unattended in a pipeline.