> ## Documentation Index
> Fetch the complete documentation index at: https://conductorone-docs-mcp-bridge-private-server.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Govern tools and toolsets

> Review discovered tools, approve and classify them, bundle them into toolsets, and bind toolsets to access profiles.

<Note>
  **Activation required.** AI access management must be enabled for your tenant before you can use it. To get started, [contact the C1 support team](mailto:support@c1.ai) for a walkthrough.
</Note>

After registering an MCP server, every tool it exposes starts life as **Unset**. This page covers reviewing those tools, approving the safe ones, classifying them, bundling them into toolsets, and binding toolsets to access profiles so end users can request them.

## Review and govern tools

Tools must be reviewed and approved before end users can request them. Use the sections below to work through the full governance workflow.

### View discovered tools

<Steps>
  <Step>
    In **Integrations > MCP servers**, open a registered server.
  </Step>

  <Step>
    Click the **Tools** tab.
  </Step>
</Steps>

The list shows every tool C1 has discovered, its current state, classification, and last-used timestamp.

### Classify a tool

C1 captures two classification axes per tool. Both are **metadata only** — there is no enforcement layer that uses them. They exist so admins can filter and reason about tools, and so enforcement can be built on top later.

| Axis       | Values                | Meaning                                      |
| :--------- | :-------------------- | :------------------------------------------- |
| **Action** | Read / Write / Delete | Whether the tool can mutate downstream state |
| **Risk**   | Sensitive / Dangerous | Severity of the worst outcome if misused     |

To classify a tool:

<Steps>
  <Step>
    Open the tool's detail panel from the **Tools** tab.
  </Step>

  <Step>
    Set **Action** and **Risk**.
  </Step>

  <Step>
    Click **Save**.
  </Step>
</Steps>

Classification can be changed at any time and does not affect approval state.

### Approve or disable a tool

A tool must be **Approved** before it can be added to any toolset.

| State        | What it means                                                             |
| :----------- | :------------------------------------------------------------------------ |
| **Unset**    | Default state for newly-discovered tools. Cannot be added to a toolset.   |
| **Approved** | Can be added to toolsets.                                                 |
| **Disabled** | Cannot be added to toolsets; existing toolsets that reference it skip it. |

To change state:

<Steps>
  <Step>
    Select one or more tools (bulk selection is supported).
  </Step>

  <Step>
    Click **Approve** or **Disable**.
  </Step>
</Steps>

### Per-tool overrides

Each approved tool has overrides that take precedence over the tenant defaults.

| Override                 | What it does                                                                                                                |
| :----------------------- | :-------------------------------------------------------------------------------------------------------------------------- |
| **Kill switch**          | Immediately blocks all calls to this tool from any client, regardless of access profile. Useful for fast incident response. |
| **Allowed client types** | Restrict this tool to a subset of personal / shared / service / ephemeral.                                                  |

### Tool lifecycle on re-sync

C1 periodically re-runs tool discovery against each registered MCP server. When the inventory changes:

* **New tool detected** — added to the list as **Unset**.
* **Existing tool changes** (description, parameters) — the change is recorded; the tool keeps its current state and classification.
* **Tool disappears** — the tool is marked **Removed**. It stays in toolsets (skipped at call time) until an admin removes it explicitly.

## Create and manage toolsets

A **toolset** is a named bundle of approved tools.

### C1-maintained toolsets

C1 ships and auto-maintains two toolsets per tenant:

* **All approved tools** — every tool in **Approved** state across the server.
* **All read tools** — every **Approved** tool with Action = Read.

Both update automatically as tools are approved or disabled. They are read-only — admins can bind them to access profiles but cannot edit their contents.

### Create a custom toolset

Custom toolsets are manually curated — they do **not** auto-populate based on classification. If you approve a new tool that you want included, you have to add it to the toolset yourself.

<Steps>
  <Step>
    Go to **AI access management > Toolsets**.
  </Step>

  <Step>
    Click **Add toolset**.
  </Step>

  <Step>
    Enter a **name**. **Optional.** Enter a **description**.
  </Step>

  <Step>
    Click **Save**.
  </Step>

  <Step>
    Click back into the toolset to add approved tools.
  </Step>
</Steps>

To edit a toolset, open it and add or remove tools. Changes propagate to any access profile the toolset is bound to.

## Bind a toolset to an access profile

AIAM uses C1's existing access profile mechanism. A toolset becomes requestable by end users only after it is bound to an access profile. There are two ways to do this.

### Option 1: From the toolset

When creating or editing a custom toolset, you can link it to an access profile directly. This is the fastest path when you've just created a custom toolset and already know which access profile it belongs to.

<Steps>
  <Step>
    Open the toolset.
  </Step>

  <Step>
    Select an existing access profile to bind it to.
  </Step>
</Steps>

The toolset's tools are now included in that profile.

### Option 2: From the access profile

You can also start from the access profile side and add toolsets as entitlements. This is required for C1-maintained toolsets and is the better path when you're setting up access profiles from scratch or adding multiple toolsets to a single profile.

<Steps>
  <Step>
    Go to **Access profiles** and either create a new profile or open an existing one.
  </Step>

  <Step>
    Add the toolset as an entitlement — both C1-maintained toolsets (**All approved tools**, **All read tools**) and custom toolsets appear as options.
  </Step>

  <Step>
    Set the access policy on the profile (auto-approve, JIT with expiry, or approval required).
  </Step>

  <Step>
    Publish the access profile to the catalog.
  </Step>
</Steps>

End users can now request the access profile and, once approved, their AI client will see the toolset's tools.