📝 refresh runtime file constructors docs

docs/api/file-sink-policy-new.md

@@ -13,7 +13,7 @@ key-word:
 
 ## File-sink-policy-new
 
-Construct a `FileSinkPolicy` value from explicit append, auto-flush, and rotation settings. This is the low-level constructor behind the public file policy shape used by direct file sinks and configured runtime file controls.
+Construct a `FileSinkPolicy` value from explicit append, auto-flush, and rotation settings. This is the low-level constructor behind the public file policy shape used by direct file sinks and higher-level runtime file-control APIs.
 
 ### Interface
 
@@ -42,7 +42,7 @@ Detailed rules explaining key parameters and behaviors
 - Omitting optional arguments uses the baseline policy: append enabled, auto-flush enabled, and no rotation.
 - This constructor simply packages the supplied settings into one public policy value.
 - It does not inspect or mutate a live file sink by itself.
-- The resulting value matches the same public shape accepted by `FileSink::set_policy(...)` and `ConfiguredLogger::file_set_policy(...)`.
+- The resulting value matches the same public shape accepted by `FileSink::set_policy(...)`, `RuntimeSink::file_set_policy(...)`, and `ConfiguredLogger::file_set_policy(...)`.
 
 ### How to Use
 

docs/api/file-sink-state-new.md

@@ -54,7 +54,7 @@ Detailed rules explaining key parameters and behaviors
 - Omitting optional arguments uses a conservative baseline snapshot: unavailable, append enabled, auto-flush enabled, no rotation, and zeroed counters.
 - This constructor simply packages the supplied fields into one public snapshot value.
 - It does not inspect a live `FileSink` by itself.
-- `FileSink::state()` and `ConfiguredLogger::file_state()` are the higher-level APIs that read these values from concrete runtime objects.
+- `FileSink::state()`, `RuntimeSink::file_state()`, and `ConfiguredLogger::file_state()` are the higher-level APIs that read these values from concrete runtime objects.
 
 ### How to Use
 
@@ -99,7 +99,7 @@ In this example, callers still use the direct constructor while making each diag
 e.g.:
 - This constructor itself does not have a normal failure mode; it only packages the provided values.
 
-- If callers want a snapshot directly from a live file sink or configured logger, `state()` or `file_state()` is the simpler API.
+- If callers want a snapshot directly from a live file sink, runtime sink, or configured logger, `state()` or `file_state()` is the simpler API.
 
 ### Notes
 

docs/api/runtime-file-state-new.md

@@ -43,8 +43,8 @@ Detailed rules explaining key parameters and behaviors
 
 - Omitting optional arguments builds a non-queued baseline around the supplied file snapshot.
 - This constructor simply packages the supplied fields into one public runtime snapshot value.
-- It does not inspect a live configured logger by itself.
+- It does not inspect a live runtime sink or configured logger by itself.
-- `ConfiguredLogger::file_runtime_state()` is the higher-level API that reads these values from concrete runtime objects.
+- `RuntimeSink::file_runtime_state()` and `ConfiguredLogger::file_runtime_state()` are the higher-level APIs that read these values from concrete runtime objects.
 
 ### How to Use
 
@@ -62,17 +62,17 @@ let snapshot = RuntimeFileState::new(
 )
 ```
 
-In this example, the runtime file snapshot is built directly without querying a live configured logger.
+In this example, the runtime file snapshot is built directly without querying a live runtime object.
 
 #### When Need Structured Runtime Diagnostics Input Before Serialization
 
 When code should prepare a typed runtime file state value for later export:
 ```moonbit
 let snapshot = RuntimeFileState::new(
-  logger.file_state(),
+  runtime.file_state(),
   queued=true,
-  pending_count=logger.pending_count(),
+  pending_count=runtime.pending_count(),
-  dropped_count=logger.dropped_count(),
+  dropped_count=runtime.dropped_count(),
 )
 ```
 
@@ -83,7 +83,7 @@ In this example, callers still use the direct constructor while making each queu
 e.g.:
 - This constructor itself does not have a normal failure mode; it only packages the provided values.
 
-- If callers want a snapshot directly from a live configured logger, `file_runtime_state()` is the simpler API.
+- If callers want a snapshot directly from a live runtime sink or configured logger, `file_runtime_state()` is the simpler API.
 
 ### Notes