Clarify and honor saved context page continuation

docs/reference/required-work-context.md

@@ -287,8 +287,10 @@ serialized response.
 Read the top-level `content` and follow its `next_command` until that selection's
 `eof`. If the whole initial selection would need paging and its decision and
 read obligations fit in one page, the opening contains those controls and
-`navigation.remaining_read_command` points to the rest. Read it and follow
-its cursor before work. Both selections pin the same immutable capture.
+`navigation.remaining_read_command` points to the rest. The page's
+`read_instruction` directs the reader to that command even when the opening
+reports EOF. Follow it and its cursor before work. Both selections pin the same
+immutable capture.
 `navigation.first_read_command` replays the opening. Use
 `navigation.recovery.command` for other fields. If the response is lost, use the
 original digest-bound capture in the wake prompt. Recovery reuses the saved
@@ -303,8 +305,9 @@ decision and its original receipt.
 
 阅读顶层 `content`，按 `next_command` 续读到当前选区的 `eof`。首读内容需要分页且决策
 和读取义务能完整放在一页时，首屏先交付这些控制信息，并给出
-`navigation.remaining_read_command`。工作前沿这条命令及其游标读完余下的首读内容。
-两个选区绑定同一份不可变捕获。`navigation.first_read_command` 可重读首屏，其他字段沿
+`navigation.remaining_read_command`。即使首屏已到 EOF，页级 `read_instruction` 仍指向
+这条命令。工作前沿这条命令及其游标读完余下的首读内容。两个选区绑定同一份不可变捕获。
+`navigation.first_read_command` 可重读首屏，其他字段沿
 `navigation.recovery.command` 读取。结果丢失时，用唤醒正文中的原文件及摘要恢复，复用原决策和回执。
 
 When the existing acceptance owner supplies complete `outcome_context`

loopx/cli_commands/quota_capture.py

@@ -749,6 +749,10 @@ def command(reference: Path) -> str:
             instruction=False, select=[], reference_file=opening, reader_argv=reader_argv)
         # An opening is useful only as a complete, independently readable
         # page. Keep the ordinary cursor when even the opening cannot fit.
+        page["read_instruction"] = (
+            "EOF ends the opening. Before work, read its content and follow navigation.remaining_read_command "
+            "and its next_command cursor through the remaining initial context. Both views pin the same capture. "
+            "EOF does not establish retention, authority or freshness.")
         page["navigation"] = navigation
         return page if len(_native_output_bytes(page)) <= output_bytes else None
     except (OSError, ValueError):
@@ -805,8 +809,7 @@ def project_saved_context_capture(args: argparse.Namespace, *, hook_session: str
         activation = native_context_activation(args)
         saved_page = bool(getattr(args, "initial_read", False) or getattr(args, "select", None))
         if saved_page and (args.format != "json" or not _context_page_output_enabled(args, activation) or args.offset != 0
-                or args.reference_file or args.selection_sha256 or args.page_bytes is not None
-                or (args.initial_read and args.select)):
+                or args.reference_file or args.selection_sha256 or (args.initial_read and args.select)):
             raise ValueError("Use JSON context-project at offset zero with native delivery or progressive --initial-read. "
                 "Choose either --initial-read or --select. Use context-read for replay or continuation.")
         caller_output_limit = getattr(args, "max_page_output_bytes", None)

tests/control_plane/test_saved_context_project.py

@@ -1155,6 +1155,9 @@ def follow(command):
     assert "without reading an older baseline" in navigation["instruction"]
     assert "Start with first_read_command" not in navigation["instruction"]
     assert ("remaining_read_command" in navigation) == staged
+    assert ("navigation.remaining_read_command" in first["read_instruction"]) == staged
+    if staged:
+        assert "Before work" in first["read_instruction"]
     assert ("next_command" in first) is paged
     if page_bytes is not None:
         for command in (navigation["first_read_command"], navigation.get("remaining_read_command"),