Hide runtime tree from feature APIs

[tiye]

Summary

• replace the exposed runtime_tree_host plumbing with a focused component_runtime capability
• make feature panel and view APIs accept typed models only, without taking or returning list<state_entry>
• add run_runtime_render so browser and test hosts thread the runtime frame in one place
• run feature components and the app shell sequentially against the same runtime capability, preventing nested runner write-back from overwriting child store updates
• remove unused tree-carrying component tuple helpers and one-line panel wrappers
• stop the Lab parent view from inspecting child draft stores for summary chips
• move raw component-state inspection helpers into test support
• document the runtime ownership boundary and parent/child state rule

Validation

• yarn test:koka: 55 / 55 pass
• yarn build
• browser regression: Todo editor draft survives Todo → Lab → Todo; Lab draft and expanded state survive Lab → Todo → Lab; no browser warnings/errors

[Copilot]

Pull request overview

This PR refactors runtime-tree plumbing so feature/panel/view APIs no longer accept or return list<state_entry>, and instead operate against a narrower component_runtime capability owned at the app boundary via runtime_frame. It centralizes render transitions with run_runtime_render(...) and enforces sequential runner execution to avoid nested runtime write-back overwriting child store updates.

Changes:

• Replace runtime_tree_host with component_runtime and update component runners to read/write component state entries via the effect.
• Introduce run_runtime_render(...) to thread runtime_frame through rendering in one place (browser host + test harnesses).
• Remove parent view “inspection” of child local stores; move raw runtime-tree inspection into test-only helpers and add regression coverage.

Reviewed changes

Copilot reviewed 21 out of 21 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
README.md	Documents the new runtime ownership boundary and parent/child local-state rule.
PLAN.md	Updates the plan notes to match the new runtime threading and sequencing approach.
explore/react/state.kk	Introduces component_runtime, updates runner plumbing, removes tree-carrying helpers.
demo/view.kk	Refactors app rendering to rely on component_runtime instead of explicit runtime-tree args/returns.
demo/todopanel.kk	Updates Todo panel runner to use run_stateful_component under component_runtime.
demo/todo/state.kk	Removes runtime-tree snapshot helpers in favor of current-store APIs.
demo/tests/todocases.kk	Switches Todo tests to test-support inspection helpers.
demo/tests/support.kk	Adds test-only inspection helpers and refactors harness rendering via run_runtime_render.
demo/tests/labcases.kk	Updates Lab tests to use run_runtime_render + inspection helpers.
demo/tests/basics.kk	Updates registry tests for the new render threading; adds app runtime regression test.
demo/tests/appharness.kk	Refactors app harness rendering through run_runtime_render.
demo/tests.kk	Registers the new regression test case.
demo/runtimeframe.kk	Adds run_runtime_render(...) to centralize render transitions at the app boundary.
demo/layout.kk	Refactors layout/stage routing to run panels sequentially under component_runtime.
demo/labpanel.kk	Updates Lab panel runner to avoid explicit runtime-tree plumbing.
demo/lab/view.kk	Removes parent view aggregation that depended on reading child local store state.
demo/lab/state.kk	Removes runtime-tree read helpers in favor of current-store APIs.
demo/effectspanel.kk	Updates Effects panel runner to use run_stateful_component under component_runtime.
demo/dialogs.kk	Moves dialog runtime reconciliation into component_runtime rather than returning updated trees.
boilerplate/browserapp.kk	Threads runtime via run_runtime_render in commit/boot paths to keep runtime ownership centralized.
Agents.md	Documents the updated rules for runtime-tree exposure and runner sequencing.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.