# Managing tools

Once your APIs are imported into AgentLink, each endpoint becomes a **Tool** that can be managed from the **Tools** screen in the dashboard.

From this screen, you can view all available tools, add new ones, adjust their configuration, control visibility, and analyze how they’re being used by AI agents.

![tools-[page]](/assets/tools-page.12de8c05b05da1eacde977cd9501af62d8db8a4df1fad6ab3694b3d09ec15ea6.b925d42d.png)

## Tools screen overview

The **Tools** screen provides a centralized view of every tool defined in your MCP server.

Each row represents one tool and typically shows:

| Field | Description |
|  --- | --- |
| **Name** | The tool name (from your OpenAPI or GraphQL schema). |
| **Description** | A summary of what the tool does. |
| **Status** | Indicates whether the tool is enabled or disabled. |
| **Auth Type** | Whether authentication is required (e.g., JWT, API Key, None). |


You can **search**, **filter**, or **sort** tools by name, category, or usage level.

## Adding new tools

You can add new tools at any time to make additional APIs available to AI agents.

To add a tool:

1. Click the **Add Tool** button in the top-right corner of the Tools screen.
2. The **Import Tools** dialog will appear.
3. Upload either:
  - An **OpenAPI** specification (`.yaml` or `.json`), or
  - A **GraphQL** schema file (`.graphql` or `.gql`)
4. AgentLink will parse your file, extract the available endpoints or operations, and create new tools automatically.
5. Once the import completes, you’ll see your new tools listed on the Tools screen, ready to review and configure.


You can then edit, enable, or disable them just like any other tool.

Tip: Re-import your API specification whenever your product’s endpoints change — AgentLink will update the tool definitions automatically.

## Enabling or disabling tools

Each tool can be enabled or disabled individually.

- **Enabled** tools are visible and callable by AI agents.
- **Disabled** tools remain defined but are hidden and cannot be invoked.


To toggle a tool:

1. Locate it in the Tools list.
2. Use the switch in the **Status** column to enable or disable it.
3. Changes take effect immediately for all connected AI agents.


Disabling tools is useful when endpoints are in development or temporarily unavailable.

## Editing tool details

You can edit tool metadata to improve accuracy or clarity for AI agents.

Editable fields include:

- **Name** (for example, rename `post_report` to `create_expense_report`)
- **Description** (clarify what the tool does and when it should be used)
- **Authentication type** (switch between “None”, “JWT”, or “API Key”)
- **Method type** (switch between "POST", "DELETE", "GET", etc)
- **API path** (for example, "/api/expenses")


To edit a tool:

1. Click anywhere on the tool’s row in the list.
2. An **Edit Tool** drawer will appear from the side.
3. Update the desired fields and click **Save**.


All changes are applied immediately once saved.

## Deleting a tool

To permanently remove a tool:

1. Click the **trash** icon on the right side of the tool’s row.
2. Confirm when prompted.


Deleting a tool removes it completely from your MCP server — AI agents will no longer be able to discover or call it.

Caution: This action cannot be undone. Make sure the tool is not being used by active agents before deleting it.

## Tool analytics

Each tool includes analytics to help you understand usage patterns:

| Metric | Description |
|  --- | --- |
| **Executions** | Total number of times the tool was called. |
| **AI Platforms** | Breakdown by ChatGPT, Claude, Gemini, or custom MCPs. |
| **Policy Actions** | Number of times the tool was allowed, denied, or required approval. |
| **Response Time** | Average duration for successful executions. |
| **Last Used** | Timestamp of the most recent execution. |


You can view these insights directly from the Tools screen or explore deeper metrics in the **Overview Dashboard**.

## Securing tool access

Even though tools are imported automatically, their access permissions are governed by:

- [Access Control](/agent-link/access-control/overview) — roles and JWT claim rules
- [Policies & Guardrails](/agent-link/policies/overview) — conditional enforcement
- [Data Protection](/agent-link/data-protection/overview) — masking and redaction rules


If a tool handles sensitive data, ensure it’s properly protected before enabling it.

## Best practices

- Keep your tool list organized with consistent naming and categorization.
- Regularly review analytics to identify high-value tools.
- Disable or delete deprecated or unused tools.
- Provide detailed descriptions to help AI agents understand context.
- Re-import your API specification when new endpoints are released (also avilable from API)


## Related topics

- [About Tools](/agent-link/tools/about-tools)
- [Access Control](/agent-link/access-control/overview)
- [Overview Dashboard](/agent-link/overview/dashboard)
- [Policies & Guardrails](/agent-link/policies/overview)