feat: add silent subagent handoff mode

[Ayleovelle]

Fixes #8181.

This PR adds a silent SubAgent handoff mode so the main agent can delegate private intermediate work to a SubAgent, consume the SubAgent result, and still answer the user without exposing the internal handoff tool call/result.

Modifications / 改动点

• Add mode: normal | silent support to SubAgent handoff tools and allow each SubAgent to define a default_handoff_mode.

• Hide silent SubAgent handoff tool_call / tool_call_result events from user-facing responses while preserving the tool result in the main agent context.

• Prevent silent SubAgents from directly calling send_message_to_user, and keep silent handoff messages out of persisted checkpoints.

• Add runtime logs that distinguish 子代理静默调用=开启 and 子代理静默调用=未开启, including whether the mode came from default config or an explicit tool argument.

• Expose the WebUI toggle as 子代理静默调用, with updated zh-CN / en-US / ru-RU translations and config API compatibility.

• Add regression tests for silent/default handoff behavior, checkpoint persistence filtering, logging, and SubAgent config loading.

• This is NOT a breaking change. / 这不是一个破坏性变更。

Screenshots or Test Results / 运行截图或测试结果

Dark mode WebUI screenshot was captured locally at data/temp/subagent-silent-toggle-dark.png, showing the 子代理静默调用 toggle enabled.

Local deployment verification:

• Started backend at http://localhost:6185 and dashboard at http://localhost:3000.
• Configured an OpenAI-compatible provider, fetched the model list, selected gpt-5.5, and passed provider health check.
• WebChat E2E verification: a silent SubAgent returned the expected final token while SSE exposed no visible tool_call / tool_call_result events.
• Runtime logs now show the difference between silent and normal handoff modes, for example 子代理静默调用=开启 vs 子代理静默调用=未开启.

Checks:

.\.venv\Scripts\ruff.exe check astrbot/core/agent/handoff.py astrbot/core/agent/message.py astrbot/core/agent/runners/tool_loop_agent_runner.py astrbot/core/astr_agent_tool_exec.py astrbot/core/subagent_orchestrator.py astrbot/dashboard/routes/subagent.py tests/test_conversation_checkpoint.py tests/test_tool_loop_agent_runner.py tests/unit/test_astr_agent_tool_exec.py tests/unit/test_subagent_orchestrator.py
All checks passed!

.\.venv\Scripts\python.exe -m pytest tests/test_tool_loop_agent_runner.py tests/unit/test_astr_agent_tool_exec.py tests/test_conversation_checkpoint.py tests/unit/test_subagent_orchestrator.py -q
67 passed

corepack pnpm run typecheck
vue-tsc --noEmit passed

---

Checklist / 检查清单

• If there are new features added in the PR, I have discussed it with the authors through issues/emails, etc. / 如果 PR 中有新加入的功能，已经通过 Issue / 邮件等方式和作者讨论过。
• My changes have been well-tested, and "Verification Steps" and "Screenshots" have been provided above. / 我的更改经过了良好的测试，并已在上方提供了“验证步骤”和“运行截图”。
• I have ensured that no new dependencies are introduced, OR if new dependencies are introduced, they have been added to the appropriate locations in requirements.txt and pyproject.toml. / 我确保没有引入新依赖库，或者引入了新依赖库的同时将其添加到 requirements.txt 和 pyproject.toml 文件相应位置。
• My changes do not introduce malicious code. / 我的更改没有引入恶意代码。

[sourcery-ai]

Sorry @Ayleovelle, your pull request is larger than the review limit of 150000 diff characters

[Ayleovelle]

[gemini-code-assist]

Code Review

This pull request introduces a 'silent' handoff mode for subagents, enabling private task execution where tool calls and raw results are hidden from the end user. Key changes include updating the HandoffTool schema, implementing message filtering during serialization to exclude private tool interactions, and enhancing the FunctionToolExecutor to handle silent mode resolution and toolset restriction. Additionally, the subagent orchestrator and dashboard UI were updated to support this configuration, supported by new unit and integration tests. I have no feedback to provide.