[Bug]: Agent loses knowledge base and command calling functionality after re-adding

[xyz66882]

Note

This issue was translated by Claude.

Issue Checklist

• I understand that issues are for feedback and problem solving, not for complaining in the comment section, and will provide as much information as possible to help solve the problem.
• My issue is not listed in the FAQ.
• I've looked at pinned issues and searched for existing Open Issues, Closed Issues, and Discussions, no similar issue or discussion was found.
• I've filled in short, clear headings so that developers can quickly identify a rough idea of what to expect when flipping through the list of issues. And not "a suggestion", "stuck", etc.
• I've confirmed that I am using the latest version of Cherry Studio.

Platform

Windows

Version

v1.6.2

Bug Description

When an agent's model settings are modified (e.g., context quantity adjustment), deleting and re-adding the agent causes the following critical functions to fail:

1. Knowledge Base Invocation Failure:

   • After re-adding the agent, the input box cannot properly invoke knowledge base functionality.
   • Users cannot access or use relevant knowledge base content in conversations.

2. MCP (Multi-Parameter Configuration) Invocation Failure:

   • After modifying model settings, MCP functionality does not respond properly.
   • Users cannot achieve expected functional interactions through MCP.

3. Slash Command (/command) Invocation Failure:

   • After re-adding the agent, slash commands (such as /command) cannot be correctly parsed or executed.
   • Users cannot trigger predefined functions or workflows through slash commands.

Steps To Reproduce

1. Prerequisites: Ensure that an agent A is already running normally in the system, and this agent has successfully linked to a knowledge base, MCP functionality, and supports slash command invocation.
2. Step 1: Modify agent model settings. Navigate to agent A's configuration page and modify its model settings, such as adjusting the "Context Length" parameter or selecting a different model version. Save these modifications.
3. Step 2: Delete the agent. Delete agent A from the home page list.
4. Step 3: Re-add agent A. Add agent A again, which should be the same as the original agent A (or at least its name/ID matches it).
5. Step 4: Attempt to invoke functions. Navigate to the conversation interface of the re-added agent A on the home page.
   • Try to use the knowledge base-related button in the input box and observe whether the knowledge base invocation button can be displayed.
   • Try to invoke MCP functionality (operate according to the specific usage method of MCP).
   • Enter the / character in the input box and observe whether the expected command list can pop up.

Expected Behavior

Expected Behavior:
After executing the reproduction steps, the re-added agent A should be able to properly invoke its associated knowledge base and MCP in the conversation interface, and when entering the / character, it should correctly display and use the slash command functionality. The deletion and re-addition operations of the agent should not cause the loss or failure of its core functional configurations.

Observed Behavior:
After executing the reproduction steps, when users operate in the conversation interface of the re-added agent A:

• The input box cannot recognize or invoke any associated knowledge base content.
• The input box cannot invoke or link to any MCP functionality.
• When entering the / character in the input box, MCP and knowledge base cannot be invoked.

Relevant Log Output

Additional Context

Reference video: #10398

---

Original Content

Issue Checklist

• I understand that issues are for feedback and problem solving, not for complaining in the comment section, and will provide as much information as possible to help solve the problem.
• My issue is not listed in the FAQ.
• I've looked at pinned issues and searched for existing Open Issues, Closed Issues, and Discussions, no similar issue or discussion was found.
• I've filled in short, clear headings so that developers can quickly identify a rough idea of what to expect when flipping through the list of issues. And not "a suggestion", "stuck", etc.
• I've confirmed that I am using the latest version of Cherry Studio.

Platform

Windows

Version

v1.6.2

Bug Description

当智能体在模型设置中进行修改(例如上下文数量调整)后,删除并重新添加智能体会导致以下关键功能失效:

1. 知识库调用失效:

   • 在重新添加智能体后,输入框无法正常调用知识库功能。
   • 用户在对话中无法访问或使用相关的知识库内容。

2. MPC(多参数配置)调用失效:

   • 修改模型设置后,MPC功能无法正常响应。
   • 用户无法通过MPC实现预期的功能交互。

3. 斜杆命令(/命令)调用失效:

   • 在重新添加智能体后,斜杳命令(如/命令)无法被正确解析或执行。
   • 用户无法通过斜杳命令触发预定的功能或工作流。

Steps To Reproduce

1. 前置条件: 确保系统中已存在一个正常运行的智能体A,并且该智能体已成功关联了知识库、MPC功能,并支持斜杠命令调用。
2. 步骤一:修改智能体模型设置。 导航至智能体A的配置页面,对其模型设置进行修改,例如调整"上下文数量(Context Length)"参数,或选择不同的模型版本。保存这些修改。
3. 步骤二:删除智能体。 从首页列表中删除智能体A。
4. 步骤三:重新添加智能体A。 再次添加智能体A,其应与原智能体A(或至少其名称/ID与其匹配)相同。
5. 步骤四:尝试调用功能。 导航至首页重新添加的智能体A的对话界面。
   • 在输入框中尝试使用知识库相关按钮,观察是否能显示调用知识库按钮。
   • 尝试调用MPC功能(根据MPC的具体使用方式进行操作)。
   • 在输入框中输入 / 字符,观察是否能弹出预期的命令列表。

Expected Behavior

预期行为:
在执行复现步骤后,重新添加的智能体A在对话界面中应能正常调用其关联的知识库和MPC,并且输入 / 字符时应能正确弹出和使用斜杠命令功能。智能体的删除和重新添加操作不应导致其核心功能的配置丢失或失效。

实际行为(Observed Behavior):
在执行复现步骤后,当用户在重新添加的智能体A的对话界面中进行操作时:

• 输入框无法识别或调用任何已关联的知识库内容。
• 输入框无法调用或关联任何MPC功能。
• 在输入框中输入 / 字符时,无法调用mcp和知识库。

Relevant Log Output

Additional Context

参考视频:#10398

[xyz66882]

@EurFelux

[EurFelux]

Cannot reproduce

[EurFelux]

Note

This issue/comment was translated by Claude.

The conditions for the knowledge base option to appear are as follows:

1. The model supports function calling and the assistant selects function calling mode
2. The assistant selects prompt calling mode

Not hiding it in the shortcut panel is a bug.

---

Original Content

知识库选项出现的条件如下：

1. 模型支持函数调用且助手选择函数调用模式
2. 助手选择提示词调用模式

在快捷面板里没有隐藏是一个bug。

[EurFelux]

Note

This issue/comment was translated by Claude.

Previously, there was a bug in the assistant settings that allowed the calling mode to have three states: prompt, function, undefined. Your situation might be caused by some error that made this setting become undefined.

---

Original Content

之前助手设置有一个bug，即允许调用模式为prompt，function，undefined三种状态，你的情况可能是由于某些错误导致这个设置变成undefined了。

[xyz66882]

Note

This issue/comment was translated by Claude.

That seems to be the case, the bug is gone after I deleted and re-added it.

---

Original Content

似乎是这样的，我重新删除添加没有这个bug了。