Self-evolving Context Database for AI Agents. Unify Agent Memory, Knowledge RAG and Skills.
Python agent-contextagent-memoryagent-pluginsagentic-ragcontext-databasedsh-pluginknowledge-basememorymemory-managementself-evolvingvector-database
feat(bot): bound concurrent subagents (#4614)
wutongyuoncea8380147a23d
diff --git a/bot/README.md b/bot/README.mdindex 4d93dc529..8b13d44e0 100644--- a/bot/README.md+++ b/bot/README.md@@ -315,6 +315,7 @@ Restart `vikingbot gateway` after changing the configuration.| `bot.agents.max_tool_iterations` | `50` | Maximum tool iterations in one turn || `bot.agents.memory_window` | `50` | Local history window and session commit message threshold || `bot.agents.subagent_enabled` | `true` | Whether to expose the `spawn` tool |+| `bot.agents.subagent_max_concurrency` | `4` | Maximum number of background subagents running at once || `bot.gateway.host` | `127.0.0.1` | Gateway listen address || `bot.gateway.port` | `18790` | Gateway listen port || `bot.sandbox.backend` | `direct` | Execution backend |diff --git a/bot/README_CN.md b/bot/README_CN.mdindex 50b588bd7..164961ae1 100644--- a/bot/README_CN.md+++ b/bot/README_CN.md@@ -312,6 +312,7 @@ export OPENVIKING_CONFIG_FILE=/path/to/ov.conf| `bot.agents.max_tool_iterations` | `50` | 单轮最大工具迭代数 || `bot.agents.memory_window` | `50` | 本地历史窗口和会话提交消息阈值 || `bot.agents.subagent_enabled` | `true` | 是否提供 `spawn` 工具 |+| `bot.agents.subagent_max_concurrency` | `4` | 同时运行的后台子 Agent 数量上限 || `bot.gateway.host` | `127.0.0.1` | Gateway 监听地址 || `bot.gateway.port` | `18790` | Gateway 监听端口 || `bot.sandbox.backend` | `direct` | 执行后端 |diff --git a/bot/tests/test_subagent_skills_context.py b/bot/tests/test_subagent_skills_context.pyindex 7062739a1..638619806 100644--- a/bot/tests/test_subagent_skills_context.py+++ b/bot/tests/test_subagent_skills_context.py@@ -2,6 +2,7 @@# SPDX-License-Identifier: AGPL-3.0"""Regression tests for subagent prompt skill loading."""+import asyncioimport sysfrom pathlib import Pathfrom types import SimpleNamespace@@ -122,3 +123,37 @@ Session-loaded instruction.assert "### Skill: session-skill" in promptassert "Session-loaded instruction." in promptassert "Global-loaded instruction." not in prompt+++async def test_subagent_spawn_rejects_tasks_above_concurrency_limit(tmp_path, monkeypatch):+ blocker = asyncio.Event()+ manager = SubagentManager(+ provider=SimpleNamespace(get_default_model=lambda: "fake-model"),+ workspace=tmp_path,+ bus=MessageBus(),+ config=SimpleNamespace(agents=SimpleNamespace(subagent_max_concurrency=1)),+ )++ async def blocked(*args, **kwargs):+ await blocker.wait()++ monkeypatch.setattr(manager, "_run_subagent", blocked)+ session_key = SimpleNamespace()++ first = await manager.spawn("first", session_key)+ second = await manager.spawn("second", session_key)++ assert "started" in first+ assert second == "Error: Subagent concurrency limit reached (1)"+ assert manager.get_running_count() == 1++ tasks = list(manager._running_tasks.values())+ blocker.set()+ await asyncio.gather(*tasks, return_exceptions=True)+ await asyncio.sleep(0)++ third = await manager.spawn("third", session_key)++ assert "started" in third+ await asyncio.gather(*manager._running_tasks.values(), return_exceptions=True)diff --git a/bot/vikingbot/agent/subagent.py b/bot/vikingbot/agent/subagent.pyindex 5c2f5432c..188ac4737 100644--- a/bot/vikingbot/agent/subagent.py+++ b/bot/vikingbot/agent/subagent.py@@ -49,6 +49,9 @@ class SubagentManager:self.temperature = temperatureself.sandbox_manager = sandbox_managerself._running_tasks: dict[str, asyncio.Task[None]] = {}+ self._max_concurrency = getattr(+ getattr(config, "agents", None), "subagent_max_concurrency", 4+ )async def spawn(self,@@ -70,6 +73,9 @@ class SubagentManager:Returns:Status message indicating the subagent was started."""+ if len(self._running_tasks) >= self._max_concurrency:+ return f"Error: Subagent concurrency limit reached ({self._max_concurrency})"+task_id = str(uuid.uuid4())[:8]display_label = label or task[:30] + ("..." if len(task) > 30 else "")diff --git a/bot/vikingbot/config/schema.py b/bot/vikingbot/config/schema.pyindex 49724f701..743bb2ccd 100644--- a/bot/vikingbot/config/schema.py+++ b/bot/vikingbot/config/schema.py@@ -469,6 +469,11 @@ class AgentsConfig(BaseModel):default=True,description="Enable the spawn tool so the main agent can start background subagents.",)+ subagent_max_concurrency: int = Field(+ default=4,+ ge=1,+ description="Maximum number of background subagents running at once.",+ )session_context_enabled: bool = Truesession_context_token_budget: int = 3000commit_token_threshold: int = 200000