studio: agent window, a chat with a model that edits the scene

✦ Agent in the Studio's top bar opens a floating chat window. The model
(any OpenAI-compatible endpoint with tool calling; default B4L's inference
server, https://inference.p0.contact/v1) gets the 15 MCP tools plus two page
tools (select_object, set_view) and the Studio's state (selection, level,
rotation) in its system prompt. Answers and reasoning stream in; tool calls
show as rows with arguments and results; object ids link to the object;
Stop ends a turn; Undo takes the agent's edits back. The window moves by
its title bar and resizes at its top-left corner.

- studio/agent.js: shared by both front ends (waits for studio.js's
  isokit:studio event); MCP over /mcp or isokit.mcp in the web Studio
- isokit studio: /api/agent/{config,models,chat} relay the calls on their
  own threads (ureq, chunked and flushed per read), so the key stays on the
  machine and the endpoint needs no CORS; settings from ISOKIT_LLM_URL /
  _KEY / _MODEL or agent.json in the user's config folder
- web Studio: calls the endpoint from the page with a key kept in the
  browser; build_web.py --agent-url names a same-site relay as default; the
  deploy builds with /llm/v1, which deploy/Caddyfile forwards to the
  inference server (its nginx refuses the CORS preflight)
- tools/mock_llm.py + tools/agent_checks.py: the agent window checks in both
  check_studio.py and check_web.py

tools/check_studio.py

@@ -1,10 +1,18 @@
-"""Functional test of the isokit Studio (GUI) and its MCP endpoint, headless Edge."""
+"""Functional test of the isokit Studio (GUI), its MCP endpoint and the agent
+window, headless Edge. The agent window runs on a mock model this script starts:
+
+    ISOKIT_LLM_URL=http://127.0.0.1:7932/v1 ISOKIT_LLM_KEY=test-key ISOKIT_CONFIG_DIR=<scratch> \
+        isokit studio <copy of examples/demo> --port 7890 --no-open
+    uv run --with playwright python tools/check_studio.py <screenshot dir> http://127.0.0.1:7890
+"""
 import json
 import sys
 import urllib.request
 from pathlib import Path
 from playwright.sync_api import sync_playwright
 
+import agent_checks
+
 BASE = sys.argv[2] if len(sys.argv) > 2 else "http://127.0.0.1:7878"
 OUT = Path(sys.argv[1])
 OUT.mkdir(parents=True, exist_ok=True)
@@ -36,6 +44,7 @@ def obj(id_):
     return next((o for o in doc()["objects"] if o["id"] == id_), None)
 
 
+mock = agent_checks.start_mock()
 with sync_playwright() as p:
     b = p.chromium.launch(channel="msedge", headless=True)
     pg = b.new_page(viewport={"width": 1500, "height": 950})
@@ -140,8 +149,17 @@ with sync_playwright() as p:
     pg.locator('[data-a="rotr"]').click()
     pg.wait_for_timeout(600)
     pg.screenshot(path=str(OUT / "studio-gf.jpg"), type="jpeg", quality=80)
+    # the agent window, on the mock model (the Studio must run with ISOKIT_LLM_URL=<mock>/v1, see the docstring)
+    cfg = get("/api/agent/config")
+    if agent_checks.MOCK not in cfg["url"]:
+        check(False, "agent window checks need the Studio started with ISOKIT_LLM_URL=" + agent_checks.MOCK + "/v1 (found " + cfg["url"] + ")")
+    else:
+        check(not cfg.get("key") and cfg["has_key"] and cfg["key_from"] == "env", "/api/agent/config says a key is set, without the key")
+        pg.locator('.iso-seg button[data-level="site"]').click()
+        agent_checks.run(pg, lambda: doc()["objects"], lambda n: pg.screenshot(path=str(OUT / n), type="jpeg", quality=80), check, can_set_key=False)
     check(not errs, f"no script errors {errs}")
     b.close()
 
+mock.terminate()
 print("FAILED:" if fails else "ALL PASSED", fails)
 sys.exit(1 if fails else 0)