feat(@projects/@claire): wire location-transparent routing into per-turn workflow

Orchestrator CLAUDE.md now instructs Claire, per turn, to call resolve_host
with the signals it extracts (explicit_host / capability_needs /
session_uuid|task_id); when the decision is NOT this node, surface that the
work belongs on that host's Claire and hand it off. Decision layer of
location-transparent Claire (13764f2f) is now live in orchestrator behavior;
cross-host execution/proxy remains the follow-up. resolve_host added to the
Plan tools list.
(manual commit via ALLOW_COMMIT — autocommit LLM still down on claire)

src/claire/orchestrator/bootstrap.py

@@ -73,7 +73,20 @@ Never redo finished work or silently drop a pending review.
 ## Per-turn workflow
 
 1. Read the user's request (everything after `[turn:<id>]`).
-2. Use the Claire MCP tools as needed:
+2. **Route the turn (location transparency).** It should not matter which
+   `[<host>] claire` the user is talking to. When the request concerns work
+   that might live on another host, call `resolve_host(...)` with the signals
+   you extract from the turn:
+   - `explicit_host` — the user named a host ("on apricot", "@black");
+   - `capability_needs` — a host-specific resource the work needs (e.g.
+     `["media"]`, `["gpu"]`, `["mount:…"]`);
+   - `session_uuid` / `task_id` — an existing session/task it is about.
+   It returns `{host, reason, candidates}`. If `host` is NOT this node, the
+   work belongs on that host's Claire — say so plainly in your reply (e.g.
+   "that runs on `[apricot] claire`") and surface the hand-off as the action.
+   If it returns this node, or there is no host-specific signal, handle it
+   here as normal. Skip routing for trivially-local requests (status, chat).
+3. Use the Claire MCP tools as needed:
    - **Read**: `list_recent_events`, `search_chat_messages`, `get_session`,
      `list_fleet` (snapshot of every active agent's current task/state).
    - **Act**: `create_project`, `add_task`, `create_assignment`,
@@ -81,12 +94,12 @@ Never redo finished work or silently drop a pending review.
    - **PM**: `create_org`, `create_person`, `create_epic`, `archive_epic`,
      `create_tag`, `transition_task_state`, `tag_task`, `untag_task`,
      `set_task_owner`, `set_task_type`, `set_task_meta`.
-   - **Plan**: `summarize_project`, `suggest_assignments`.
+   - **Plan**: `summarize_project`, `suggest_assignments`, `resolve_host`.
    - **Reference**: `status`, `list_tasks`, `help`.
-3. **Always call `report_status`** once per turn with your own
+4. **Always call `report_status`** once per turn with your own
    `session_uuid` (look in `$CLAUDE_CODE_SESSION_ID`) + a one-line summary
    so the fleet view stays current.
-4. When done, call `submit_chat_reply(body=<your reply>, turn_id=<id>)`.
+5. When done, call `submit_chat_reply(body=<your reply>, turn_id=<id>)`.
    This is REQUIRED — without it, the user sees nothing.
 
 Exactly one `submit_chat_reply` call per user turn.