Lantern’s app writer should generate a copyable AI authoring prompt from the exact reviewed or planned app context. The prompt should include the app purpose, activity type, allowed GatewayApp methods, declared capabilities, grading mode, content schema, preview expectations, accessibility requirements, and explicit forbidden behaviors such as raw LMS access, arbitrary outbound HTTP, direct storage, package installs, or unreviewed backend code. This should be generated from the same app plan/capability data used by validation so the prompt cannot drift from the actual contract. The goal is to make AI-assisted app creation more reliable without exposing private product strategy, deployment details, secrets, or internal ops context.
Lantern’s app writer should generate a copyable AI authoring prompt from the exact reviewed or planned app context. The prompt should include the app purpose, activity type, allowed
GatewayAppmethods, declared capabilities, grading mode, content schema, preview expectations, accessibility requirements, and explicit forbidden behaviors such as raw LMS access, arbitrary outbound HTTP, direct storage, package installs, or unreviewed backend code. This should be generated from the same app plan/capability data used by validation so the prompt cannot drift from the actual contract. The goal is to make AI-assisted app creation more reliable without exposing private product strategy, deployment details, secrets, or internal ops context.