rigor-dry-schema

Recognises dry-schema declarations and publishes a per-schema typed-key table as a cross-plugin fact (:dry_schema_table) that rigor-dry-validation consumes for typed-payload synthesis. Like rigor-dry-types, it is a foundation plugin: no diagnostics of its own, no config keys.

It ships bundled in rigortype. Activate it (pair with rigor-dry-types to resolve Types::* aliases inside predicate arguments):

plugins:
  - rigor-dry-types     # optional: resolves Types::Email etc.
  - rigor-dry-schema

What it recognises

NewUserSchema = Dry::Schema.Params do
  required(:email).filled(:string)
  required(:age).value(:integer)
  required(:tags).each(:string)
  optional(:nickname).maybe(:string)
end

• required / optional keys, with the predicate’s type symbol mapped to a Ruby class (:string → String, :integer → Integer, :decimal → BigDecimal, :bool → TrueClass, …).
• each(:T) marks the key as a list (list: true); filled / value / maybe are scalar (list: false).
• value(Types::Email) resolves through the :dry_type_aliases fact when rigor-dry-types is loaded; without it (or for an unknown reference) the row drops from the table rather than mislead downstream consumers.
• Top-level (Foo = Dry::Schema.Params { … }) and class-level (class Bar; SCHEMA = …; end → "Bar::SCHEMA") declarations, across .Params / .JSON / .define.

No diagnostics, no config

The plugin only publishes the schema table for other plugins to consume; it surfaces no diagnostics and takes no config keys. (A future slice adds dry-schema.unknown-predicate / unknown-type info diagnostics and typed result.to_h synthesis.)

Plugin internals

The prepare(services) scan, the published :dry_schema_table shape, and the slice floor/ceiling are documented in the plugin’s README. To write a plugin, see examples/ and the rigor-plugin-author skill.