Validate and publish a rule

Validation checks technical correctness. Test computes private results. Activation permits the exact source version and derived definition to be used against business data.

Prompt

In {{workspace_id}} / {{application_id}}, verify source-design and publication authority and the approved Python environment.
Save {{source_text}} with python.program.save using programId {{program_id}}, version {{version}}, entryPoint build, parameterSchema {{parameter_schema}}, environment identity and current metadata/policy/schema pins. Use {{save_key}}.
Poll the returned Validation task; report diagnostics with source locations until validation succeeds. Read back the original source and sourceDigest.
Request python.program.test for this exact source. Supply builder parameters {{parameters}} separately from definitionParameters {{rule_bindings}}. Use a stable Test key and current pins.
Inspect the completed Impact artifact, including actual bindings, source/definition identity, baseline and proposed effects. Business data must remain unchanged.
When asked to activate, call python.program.publish with the exact source version, current pins and validationBuildId {{reviewed_build_id}}. Use a separate stable activation key.
Verify the activated definition's source binding and current publication. Do not run it as a side effect of activation.

Example

For docs.input-uplift, use builder parameters {} and rule bindings {"InputAmount":"12.50000000"}. Test should show 15.00000000 at the previously absent destination while an ordinary read still shows absence. Activation preserves that absence. Its explicit build identity prevents another client's later Test from changing which definition is activated.

Source registry tool · Test and build tool · Run and inspect results