Tool visibility

Choose which of a server's tools AI agents can call. Disabled tools are hidden from every client, so agents never see them in their tool list.

Configure it on the server

Go toQuilrAI consoleSettingsAI GatewayMCP Gatewayserver cardConfigureTools

You can also open Overall analytics > Settings, select one server and choose Tools. The section subtitle shows how many tools are exposed, for example 12 of 15 exposed.

Tool groups

Tools are grouped by what they do to the upstream system.

Group	What the tools do
Read-only tools	Only read data.
Write tools	Create or change data.
Destructive tools	Delete data or make irreversible changes.

The Exposed to agents bar at the top shows how many tools in each group are enabled, for example Write 3 / 4.

Per-tool controls

Column	What it does
Tool	Name and description. Click View input schema to see the inputs the tool accepts.
Calls	How many calls the tool has received.
Enabled	Off hides the tool from every client.
Require confirmation	The user must approve each call before it runs. See Human approval.
Require justification	The user must type a reason to approve. Only available with confirmation on.

Changes are drafts. Click Save settings in the footer to apply them.

Find a tool

Control	Use
Search tools	Filter by tool name or description.
All / Read-only / Write / Destructive	Show one group. Each tab shows its tool count.

Refresh the tool list

Click Refresh tools to fetch the server's current tool list. If the list is empty, the gateway hasn't fetched the tools yet; refresh, or connect OAuth first if the server needs it.

The Upstream tool changes card above the tool list watches for tools the server adds, removes or changes. See Tool change watch.

Per group and per user

To enable a tool, or require confirmation, only for some people, add a rule in Group and user rules. A rule can set each tool's Enabled, Confirmation and Justification to Inherit, On or Off for one smart group or one user.

Going further with the Policy Engine

When the Policy Engine is on for the MCP Gateway, the Tools section turns read-only and shows Controlled by Policy Engine; values saved through Edit anyway apply only if the Policy Engine is disabled (see What happens to classic settings). Tool availability then comes from two cards in Govern > Policy Engine > MCP Gateway:

Card	Stage	What a deny does
Discovery Visibility	2, Discovery	Removes the tool from the list the agent is offered.
Invocation	3, Request	Refuses the call if the agent calls the tool anyway.

Use both. An agent that cached an earlier tool list can still attempt a call, so hiding a tool alone is not enough. Edits join a shared draft and apply once you publish a revision.

Scenarios the cards support that the Tools switches cannot:

• Hide tools by what they do, across servers. Match the tool's read_only, destructive, idempotent or open_world annotations, its tags or its risk, instead of switching tools off one by one.
• Hide tools from some callers only. Combine a tool condition with smart groups, user email or agent name, for example hide write tools from one AI client.
• Hide resources and prompts. Besides tools, Discovery Visibility can deny MCP resources (by URI, template, name or MIME type) and prompts (by name), which have no switch in server settings.

hide_destructive_github_tools

·priority 850

Rule 1 - runs on discovery

WhenMCP nameisGitHub

andTool is destructiveistrue

Then

Tool call access→deny

Rule 2 - runs on request

WhenMCP nameisGitHub

andTool is destructiveistrue

Then

Tool call access→deny

Related

• Human approval - what approvers see in each AI client.
• Tool change watch - review upstream tool changes before agents see them.
• Input aliases - translate mismatched tool inputs from AI clients.
• Security guardrails - scan tool inputs and results.
• Policy Engine overview - how cards, stages and priorities work.