perf(studio): create sandbox tools in parallel (#724)

* perf(studio): create sandbox tools in parallel

* fix(frontend): align agent selector icon

Author: yaozheng-fang

docs/content/docs/framework/frontend.en.mdx

@@ -223,8 +223,8 @@ frontend-process memory, so polling and downloads continue to work when FaaS
 requests reach different instances.
 
 On the first cloud deployment, the command creates two independent CodeEnv
-Tools with the deployer's AK/SK when Tool IDs are omitted. Existing Tools can be
-selected with `--sandbox-chat-codex-tool-id` and
+Tools in parallel with the deployer's AK/SK when Tool IDs are omitted. Existing
+Tools can be selected with `--sandbox-chat-codex-tool-id` and
 `--sandbox-skill-creator-tool-id`. The Ark credential is resolved with the
 deployer's AK/SK and written directly to both Tools as model environment
 variables. Sandbox sessions inherit those Tool variables. The browser and

docs/content/docs/framework/frontend.mdx

@@ -191,7 +191,7 @@ ZIP 上传到 TOS，再调用 Skills API 创建或更新 Skill；只有用户填
 候选 Session 的 TTL 为 30 分钟，用户重新创建或离开任务时会立即清理。任务状态保存
 在 Sandbox 而非 Frontend 进程内存，因此 FaaS 请求命中不同实例时仍可继续轮询和下载。
 
-云上首次部署时，若没有指定 Tool ID，部署命令会使用部署 AK/SK 自动创建两个独立
+云上首次部署时，若没有指定 Tool ID，部署命令会使用部署 AK/SK 并行创建两个独立
 CodeEnv Tool。也可以分别使用 `--sandbox-chat-codex-tool-id` 和
 `--sandbox-skill-creator-tool-id` 指定已有 Tool。部署命令会通过 AgentKit 凭据托管
 将 Ark 凭据保存到 KMS，并把代理地址和可撤销票据绑定到两个 Tool；VeFaaS Function

frontend/README.md

@@ -289,8 +289,8 @@ memory, so polling and downloads continue to work when FaaS requests reach a
 different instance.
 
 For local Studio, run the AgentKit `credential-hosting` command and bind its
-result to both CodeEnv Tools. A cloud deployment creates both Tools when their
-IDs are omitted. Alternatively, select existing Tools with
+result to both CodeEnv Tools. A cloud deployment creates both Tools in parallel
+when their IDs are omitted. Alternatively, select existing Tools with
 `--sandbox-chat-codex-tool-id` and `--sandbox-skill-creator-tool-id`. The deploy
 command obtains the Ark key with the deployer's Volcengine credentials, stores
 it through AgentKit credential hosting, and binds only the returned ticket and

frontend/src/styles.css

@@ -247,7 +247,7 @@ body {
   border-color: hsl(8 46% 70%);
   background: hsl(8 55% 96%);
 }
-.agent-row-lead { width: 16px; height: 16px; color: hsl(var(--muted-foreground)); flex-shrink: 0; transform: translateY(3px); }
+.agent-row-lead { width: 16px; height: 16px; color: hsl(var(--muted-foreground)); flex-shrink: 0; }
 .agent-row--connected .agent-row-lead { color: hsl(142 48% 38%); }
 .agent-row-name {
   flex: 1;

frontend/tests/studioAccess.test.mjs

@@ -60,7 +60,8 @@ test("runtime authorization failures are not reported as unsupported", () => {
 });
 
 test("selected Agent icons are optically aligned with the label", () => {
-  assert.match(stylesSource, /\.agent-row-lead[^}]*transform: translateY\(3px\)/);
+  assert.match(stylesSource, /\.agent-row\s*{[^}]*align-items: center/);
+  assert.doesNotMatch(stylesSource, /\.agent-row-lead[^}]*transform:/);
   assert.match(stylesSource, /\.agent-row-chev\.open[^}]*translateY\(1px\) rotate\(90deg\)/);
 });
 

tests/cli/test_studio_deploy_target.py

@@ -12,6 +12,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import threading
 from pathlib import Path
 from types import SimpleNamespace
 from unittest.mock import Mock
@@ -294,8 +295,12 @@ def register_callback_for_user_pool_client(self, **kwargs: object) -> None:
 def test_studio_deploy_creates_distinct_sandbox_tools_when_ids_are_omitted(
     monkeypatch: pytest.MonkeyPatch,
 ) -> None:
+    monkeypatch.delenv("SANDBOX_CHAT_CODEX", raising=False)
+    monkeypatch.delenv("SANDBOX_SKILL_CREATOR", raising=False)
     created_names: list[str] = []
     credential_tool_ids: list[str] = []
+    creation_barrier = threading.Barrier(2)
+    created_names_lock = threading.Lock()
 
     class _FakeCloudAgentEngine:
         def __init__(self, **_: object) -> None:
@@ -310,8 +315,10 @@ def deploy(self, **_: object) -> SimpleNamespace:
 
     def _ensure_tool(**kwargs: object) -> str:
         name = str(kwargs["name"])
-        created_names.append(name)
-        return f"tool-{len(created_names)}"
+        creation_barrier.wait(timeout=5)
+        with created_names_lock:
+            created_names.append(name)
+        return "chat-tool" if "-chat-" in name else "skill-tool"
 
     monkeypatch.setattr(
         "veadk.cloud.cloud_agent_engine.CloudAgentEngine", _FakeCloudAgentEngine
@@ -351,12 +358,11 @@ def _ensure_tool(**kwargs: object) -> str:
 
     assert result.exit_code == 0, result.output
     assert len(created_names) == 2
-    assert created_names[0] != created_names[1]
-    assert "chat" in created_names[0]
-    assert "skill" in created_names[1]
-    assert veadk_environments["SANDBOX_CHAT_CODEX"] == "tool-1"
-    assert veadk_environments["SANDBOX_SKILL_CREATOR"] == "tool-2"
-    assert credential_tool_ids == ["tool-1", "tool-2"]
+    assert any("chat" in name for name in created_names)
+    assert any("skill" in name for name in created_names)
+    assert veadk_environments["SANDBOX_CHAT_CODEX"] == "chat-tool"
+    assert veadk_environments["SANDBOX_SKILL_CREATOR"] == "skill-tool"
+    assert credential_tool_ids == ["chat-tool", "skill-tool"]
 
 
 @pytest.mark.parametrize(

veadk/cli/cli_frontend.py

@@ -30,6 +30,7 @@
 import sys
 
 from collections.abc import Iterable
+from concurrent.futures import ThreadPoolExecutor
 from pathlib import Path
 from typing import Any
 from urllib.parse import urlparse
@@ -3925,29 +3926,40 @@ def frontend_deploy(
         studio_sandbox_tool_name,
     )
 
+    missing_sandbox_tools: dict[str, str] = {}
     for purpose, tool_id in sandbox_tool_ids.items():
         if tool_id:
             continue
         tool_name = studio_sandbox_tool_name(vefaas_app_name, purpose)
         click.echo(f"Ensuring AgentKit {purpose} CodeEnv Tool '{tool_name}'…")
-        try:
-            sandbox_tool_ids[purpose] = ensure_studio_code_env_tool(
-                name=tool_name,
-                region=region,
-                access_key=ak,
-                secret_key=sk,
-                session_token=session_token or "",
-            )
-        except Exception as error:
-            detail = _safe_exception_detail(
-                error,
-                secrets=(ak, sk, session_token),
-            )
-            raise click.ClickException(
-                f"Failed to provision the AgentKit {purpose} CodeEnv Tool. "
-                f"Underlying error:\n{detail}"
-            ) from error
-        click.echo(f"AgentKit {purpose} CodeEnv Tool is ready.")
+        missing_sandbox_tools[purpose] = tool_name
+
+    if missing_sandbox_tools:
+        with ThreadPoolExecutor(max_workers=len(missing_sandbox_tools)) as executor:
+            tool_futures = {
+                purpose: executor.submit(
+                    ensure_studio_code_env_tool,
+                    name=tool_name,
+                    region=region,
+                    access_key=ak,
+                    secret_key=sk,
+                    session_token=session_token or "",
+                )
+                for purpose, tool_name in missing_sandbox_tools.items()
+            }
+            for purpose, future in tool_futures.items():
+                try:
+                    sandbox_tool_ids[purpose] = future.result()
+                except Exception as error:
+                    detail = _safe_exception_detail(
+                        error,
+                        secrets=(ak, sk, session_token),
+                    )
+                    raise click.ClickException(
+                        f"Failed to provision the AgentKit {purpose} CodeEnv Tool. "
+                        f"Underlying error:\n{detail}"
+                    ) from error
+                click.echo(f"AgentKit {purpose} CodeEnv Tool is ready.")
 
     from veadk.cli.frontend_skill_creator import (
         ensure_skill_creator_model_credential,