docs(adr-0011): record why request-batch keeps its dead outer chain-error

crates/nexum-runtime/src/host/impls/chain.rs

@@ -79,7 +79,8 @@ impl<T: RuntimeTypes> nexum::host::chain::Host for HostState<T> {
     /// `RpcResult::Err`. This impl folds each entry independently, so a
     /// failure leaves its neighbours intact; a different host could instead
     /// short-circuit the batch, so SDK consumers match on each entry, not
-    /// on the batch call.
+    /// on the batch call. See ADR-0011's escape-hatch note for why the
+    /// WIT keeps the outer error despite it being dead here.
     async fn request_batch(
         &mut self,
         chain_id: u64,

docs/adr/0011-per-interface-typed-errors.md

@@ -29,3 +29,4 @@ This is a pre-1.0 wire break. CI rebuilds every module wasm on a world change, s
 - Modules no longer restate their identity or prefix messages; the runtime supplies both.
 - The numeric `code` and opaque `data` fields are gone. An interface that needs structured detail (the JSON-RPC code, decoded revert bytes) carries it in a typed case instead.
 - Old bindings do not interoperate with the new world. Because this predates 1.0 and CI rebuilds all module wasms per world change, that break is accepted rather than shimmed.
+- `chain-error`'s outer position on `request-batch` (`result<list<rpc-result>, chain-error>`) is structurally dead on the reference host: every failure - a denied method, a node revert, a transport fault - surfaces per-entry as `rpc-result::err`, folding the batch independently, so the outer `Err` arm is never returned. It stays in the WIT contract as an escape hatch for a host that cannot make that same choice - one whose batch transport can only fail atomically before dispatching any entry - not because this host needs it. See the doc comment on `chain::Host::request_batch` (`crates/nexum-runtime/src/host/impls/chain.rs`).