In a Symfony + API Platform project with pragmatic hexagonal architecture, what is usually recommended for write operations?

Is it better to keep:

• a dedicated API HTTP input DTO / payload representing the HTTP contract
• a separate application Command representing the application / use-case contract

or is it acceptable to use the application Command itself as the API Platform input?

More specifically:

• when should the HTTP API contract and the application contract stay as two separate objects?
• when is it reasonable to collapse them into a single Command used directly as API input?
• if we use the Command as input, does that weaken an important architectural boundary even if dependency direction still remains Infrastructure -> Application -> Domain?
• in both approaches, when is a custom Processor recommended?