diff --git a/docs/api/file-sink-policy.md b/docs/api/file-sink-policy.md
new file mode 100644
index 0000000..4b9db3b
--- /dev/null
+++ b/docs/api/file-sink-policy.md
@@ -0,0 +1,79 @@
+---
+name: file-sink-policy
+group: api
+category: sink
+update-time: 20260613
+description: Public file policy alias used by file sinks and configured runtime file controls.
+key-word:
+    - file
+    - policy
+    - alias
+    - public
+---
+
+## File-sink-policy
+
+`FileSinkPolicy` is the public policy object used to describe file append mode, auto-flush behavior, and optional rotation settings together. It is a direct alias to the file policy model shared by `FileSink`, `ConfiguredLogger`, and file policy JSON helpers.
+
+### Interface
+
+```moonbit
+pub type FileSinkPolicy = @utils.FileSinkPolicy
+```
+
+#### output
+
+- `FileSinkPolicy` - Public file policy object containing `append`, `auto_flush`, and optional `rotation` settings.
+
+### Explanation
+
+Detailed rules explaining key parameters and behaviors
+
+- This is a type alias, not a separate runtime wrapper.
+- The current policy fields are `append : Bool`, `auto_flush : Bool`, and `rotation : FileRotation?`.
+- The same policy object is returned by `FileSink::policy()` and `ConfiguredLogger::file_policy()`.
+- It is also accepted by `FileSink::set_policy(...)` and `ConfiguredLogger::file_set_policy(...)`.
+
+### How to Use
+
+Here are some specific examples provided.
+
+#### When Need One Object For Runtime File Settings
+
+When append, flush, and rotation behavior should be updated together:
+```moonbit
+let policy = FileSinkPolicy::new(
+  append=false,
+  auto_flush=true,
+  rotation=Some(file_rotation(1024 * 1024, max_backups=3)),
+)
+```
+
+In this example, the file policy can be passed around as one typed value instead of separate flags.
+
+#### When Need Runtime Policy Roundtrip
+
+When current file behavior should be read, adjusted, and written back:
+```moonbit
+let policy = logger.file_policy()
+let next = FileSinkPolicy::new(
+  append=policy.append,
+  auto_flush=false,
+  rotation=policy.rotation,
+)
+```
+
+In this example, the policy object is the handoff boundary for runtime file control.
+
+### Error Case
+
+e.g.:
+- `FileSinkPolicy` itself does not have a runtime failure mode.
+
+- If a non-file runtime sink exposes a fallback policy through file helpers, the object is still valid but does not reflect a live file handle.
+
+### Notes
+
+1. Use `FileSinkPolicy::new(...)` to construct this policy explicitly.
+
+2. Use `file_sink_policy_to_json(...)` or `stringify_file_sink_policy(...)` when the policy should be exported.
diff --git a/docs/api/file-sink-state.md b/docs/api/file-sink-state.md
new file mode 100644
index 0000000..fcf5a50
--- /dev/null
+++ b/docs/api/file-sink-state.md
@@ -0,0 +1,70 @@
+---
+name: file-sink-state
+group: api
+category: runtime
+update-time: 20260613
+description: Public file state alias used for live file-sink snapshots and runtime diagnostics.
+key-word:
+    - file
+    - state
+    - alias
+    - public
+---
+
+## File-sink-state
+
+`FileSinkState` is the public snapshot object used to describe the current state of a file sink. It is a direct alias to the file state model returned by `FileSink::state()` and configured runtime file inspection helpers.
+
+### Interface
+
+```moonbit
+pub type FileSinkState = @utils.FileSinkState
+```
+
+#### output
+
+- `FileSinkState` - Public file state snapshot containing path, availability, policy flags, optional rotation, and failure counters.
+
+### Explanation
+
+Detailed rules explaining key parameters and behaviors
+
+- This is a type alias, not a live file handle wrapper.
+- The current snapshot fields are `path`, `available`, `append`, `auto_flush`, `rotation`, `open_failures`, `write_failures`, `flush_failures`, and `rotation_failures`.
+- `FileSink::state()` returns this object directly for a concrete file sink.
+- `ConfiguredLogger::file_state()` also returns this type, including fallback snapshots for non-file runtime sinks.
+
+### How to Use
+
+Here are some specific examples provided.
+
+#### When Need A One-shot File Health Snapshot
+
+When file runtime diagnostics should be read as one object:
+```moonbit
+let state = sink.state()
+```
+
+In this example, availability, policy, and failure counters are captured together.
+
+#### When Need To Export File Diagnostics
+
+When a snapshot should be serialized for logs or support output:
+```moonbit
+println(stringify_file_sink_state(logger.file_state(), pretty=true))
+```
+
+In this example, the typed snapshot becomes readable JSON without manual field assembly.
+
+### Error Case
+
+e.g.:
+- `FileSinkState` itself does not have a runtime failure mode.
+
+- If the sink is unavailable, the snapshot still exists and reports `available=false` together with the current counters.
+
+### Notes
+
+1. Use this object for one-shot file diagnostics instead of many narrow reads.
+
+2. Use `RuntimeFileState` when queue-related file runtime context is also required.
diff --git a/docs/api/index.md b/docs/api/index.md
index 1496898..5cfca78 100644
--- a/docs/api/index.md
+++ b/docs/api/index.md
@@ -103,10 +103,13 @@ BitLogger API navigation.
 - [file-sink.md](./file-sink.md)
 - [file-rotation.md](./file-rotation.md)
 - [file-sink-policy-to-json.md](./file-sink-policy-to-json.md)
+- [file-sink-policy.md](./file-sink-policy.md)
 - [stringify-file-sink-policy.md](./stringify-file-sink-policy.md)
 - [file-sink-state-to-json.md](./file-sink-state-to-json.md)
+- [file-sink-state.md](./file-sink-state.md)
 - [stringify-file-sink-state.md](./stringify-file-sink-state.md)
 - [runtime-file-state-to-json.md](./runtime-file-state-to-json.md)
+- [runtime-file-state.md](./runtime-file-state.md)
 - [stringify-runtime-file-state.md](./stringify-runtime-file-state.md)
 - [sink-config.md](./sink-config.md)
 - [sink-config-to-json.md](./sink-config-to-json.md)
diff --git a/docs/api/runtime-file-state.md b/docs/api/runtime-file-state.md
new file mode 100644
index 0000000..7540d8d
--- /dev/null
+++ b/docs/api/runtime-file-state.md
@@ -0,0 +1,76 @@
+---
+name: runtime-file-state
+group: api
+category: runtime
+update-time: 20260613
+description: Public combined file-and-queue runtime state alias used by configured logger diagnostics.
+key-word:
+    - runtime
+    - file
+    - state
+    - public
+---
+
+## Runtime-file-state
+
+`RuntimeFileState` is the public snapshot object that combines file state with queue runtime context for configured loggers. It is a direct alias to the runtime model used by configured file diagnostics and JSON export helpers.
+
+### Interface
+
+```moonbit
+pub type RuntimeFileState = @utils.RuntimeFileState
+```
+
+#### output
+
+- `RuntimeFileState` - Public combined runtime snapshot containing a `file` snapshot plus queue metadata such as `queued`, `pending_count`, and `dropped_count`.
+
+### Explanation
+
+Detailed rules explaining key parameters and behaviors
+
+- This is a type alias, not a live runtime controller.
+- The current fields are `file : FileSinkState`, `queued : Bool`, `pending_count : Int`, and `dropped_count : Int`.
+- This type is returned by `ConfiguredLogger::file_runtime_state()` when the configured sink can expose file runtime context.
+- It is broader than `FileSinkState` because it also reports whether queue wrapping is involved and how the queue is behaving.
+
+### How to Use
+
+Here are some specific examples provided.
+
+#### When Need Combined File And Queue Diagnostics
+
+When a configured logger may wrap a file sink in a queue:
+```moonbit
+match logger.file_runtime_state() {
+  Some(snapshot) => println(snapshot.pending_count)
+  None => ()
+}
+```
+
+In this example, queue backlog can be inspected alongside the embedded file state.
+
+#### When Need A Single Exportable Runtime Snapshot
+
+When runtime support output should include both file and queue context:
+```moonbit
+match logger.file_runtime_state() {
+  Some(snapshot) => println(stringify_runtime_file_state(snapshot, pretty=true))
+  None => ()
+}
+```
+
+In this example, one typed object covers both the file snapshot and queue counters.
+
+### Error Case
+
+e.g.:
+- `RuntimeFileState` itself does not have a runtime failure mode.
+
+- If the configured logger is not file-backed, `file_runtime_state()` may return `None` instead of producing this snapshot.
+
+### Notes
+
+1. Use this type when queued-file diagnostics matter, not just raw file status.
+
+2. Use `runtime_file_state_to_json(...)` or `stringify_runtime_file_state(...)` when the snapshot should be exported.