Resources · Claude connector

Claude MuleSoft connectorWhat Claude can do in your MuleSoft account.

The Claude MuleSoft connector exposes 66 tools: 52 read, 9 write, 5 undocumented. Below: what Claude can look up across your Anypoint landscape, what it can apply, provision or delete, the connected app you need first, and the tools no source describes.

Verified Trustpilot reviews · AI, automation & growth agency

Overview

What changes for your integration team

Answering a simple question about your API estate usually means switching between Exchange, API Manager, Monitoring and governance reports. With the connector, a platform lead asks Claude in plain English: it searches the catalogue, reads the governance and traffic data behind each service, and can apply a policy or deploy a server once you agree. The consoles stay the source of truth; Claude becomes a faster way in.

Find what already exists before building it. A product team asks whether a payments API is already published. search_assets_semantic searches the portfolio by meaning, not keywords, and view_api_version_details shows the instances and lifecycle of the match.

Spot the service that hurts. Error rates rose overnight. fetch_monitoring_overview gives the org-wide snapshot, then fetch_monitoring_drill_down ranks which services or instances drive the errors or the latency.

Tighten governance without a ticket queue. Asked to add rate limiting on one API, Claude loads the options with prepare_policy_creation, shows the form through get_policy_template_form, then apply_policy_to_instance puts it in place. MuleSoft's own example says the policy applies immediately.

What the directory sheet leaves out: nothing connects until someone creates a connected app in Anypoint Platform, and many tools need an extra Anypoint permission on top. Five tools on the sheet are absent from MuleSoft's current reference. No source says whether Claude prompts before the nine write tools. And nothing runs by itself: Claude acts while you talk to it. For CloudHub actions inside scheduled scenarios instead, the MuleSoft Make integration covers that use.

Vocabulary

The vocabulary in one minute

Five words you will meet while connecting MuleSoft.

Connector
The link you set up once between Claude and an account you already have, so Claude can work in it while it answers you.
Tool
One named thing a connector lets Claude do. Claude picks the ones it needs on its own; the directory lists them by name.
Authorization
The service's own sign-in screen, where you hand Claude the access it will use. Given once per person, and revocable.
Approval
The confirmation Claude waits for before it goes ahead with something that changes your account, shown in the chat when it matters.
MCP
The shared standard connectors are built on: it is what lets an assistant like Claude talk to an outside service.
Connect

Connect MuleSoft to Claude in three steps

  1. 01

    Find MuleSoft in the list

    In Claude's settings, open Customize, then Connectors, and look for MuleSoft. On a Team or Enterprise workspace, an Owner or Primary Owner has to enable the connector first before each member can sign in.

  2. 02

    Start the connection

    Click Connect on its row. MuleSoft's guide adds one step here: enter the Client ID and Client Secret of your Anypoint connected app, then sign in to Anypoint in the window that opens. If the link breaks, use Disconnect.

  3. 03

    Read the authorization screen

    Read Anypoint's authorization screen before you approve. It belongs to the service, not to Claude, and it sets what the access covers. From then on, Claude works through the connected app, which acts on your behalf.

Tools

The 66 MuleSoft tools, grouped by what they do

MuleSoft gives Claude 66 tools: 52 that read your account, 9 that change something in it, and 5 no official source describes.

Three groups: what Claude reads, what it applies, deploys or deletes, and five names that MuleSoft's current reference does not describe. Names stay the ones Claude displays, in English.

  • 52 read
  • 9 write
  • 5 not documented

What Claude reads (52)

52 tools

Fifty-two tools that browse, search and report on your landscape without touching any resource.

check_policy_conformance

Converts a selected policy payload into a structured linked-data format and prints it in the server's logs.

When it helps
a support engineer at MuleSoft asks for the exact payload behind a policy that behaves oddly on one instance.
Watch out
MuleSoft describes the output as printed in backend logs, and says nothing more about it.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

fetch_automated_policy_template

Fetches one policy template by its identifier, with the configuration details needed to fill it in.

When it helps
before agreeing to an automated strategy, you want to see exactly which settings a header-injection template expects from you, and which ones are optional.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

fetch_cost_instance_timeseries

Returns token usage and response time over time for a single instance, and MuleSoft notes it is the only tool that gives an agent's timeline.

When it helps
finance asks why one support agent's token bill doubled in March and you need the curve, not a total.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

fetch_cost_instances

Lists token usage and optimization coverage instance by instance, across tool servers and agents, including which cost policies are not yet applied.

When it helps
you want the five most expensive instances and whether any saving policy is still switched off on them.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

fetch_cost_optimization_recommendations

Ranks the instances with the highest projected token savings from cost policies that have not been applied yet.

When it helps
the platform team has one sprint for optimisation and wants to start where the gain is largest, with a ranked list to share in the planning meeting.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

fetch_cost_overview

Gives the org-wide cost summary: token totals, tokens saved by optimisation policies and tool call counts, plus agent spend in dollars when cost data exists.

When it helps
a quarterly review needs one slide on what your AI traffic costs and what policies saved.
Watch out
the dollar cards are absent when your org has no cost data.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

fetch_governance_service_report

Returns detailed governance conformance data for a single service, rule by rule.

When it helps
an audit flagged one API and you need to know precisely which governance rules it fails before the review meeting, with enough detail to assign each fix to someone.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

fetch_governance_services

Searches, filters and counts governance conformance across all your services at once, for the big picture rather than one API.

When it helps
the CISO wants the number of customer-facing APIs still failing the PII rules this week, compared with the figure from last month.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

fetch_monitoring_drill_down

Ranks services or instances by their share of one metric: availability, reliability, average latency, request volume, error rate or policy violations.

When it helps
after an incident, you ask which services drove the errors and get them in order, worst first, to decide where the post-mortem should look.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

fetch_monitoring_instance

Pulls monitoring data for one API instance: charts for requests, error rate, latency and violations, plus its busiest endpoints and clients.

When it helps
a partner complains about timeouts and you want their instance's latency and top callers side by side.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

fetch_monitoring_overview

Shows an org-wide health snapshot: total requests, error rate, average latency, availability, reliability and compact trend charts.

When it helps
the Monday operations review starts with how the whole estate did last week, before anyone drills into a single service.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

fetch_service_facets

Computes aggregate counts and buckets, by type or provider for instance, when a summary is needed rather than full service rows.

When it helps
you want to know how many APIs, agents and LLMs each provider holds, as three numbers for a board slide.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

fetch_services

Returns concrete service entries with identifiers and names, the tool Claude uses when it needs real rows to pick from or open.

When it helps
after a summary, you ask for the actual list of agents on AWS Bedrock to review them one by one.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

get_business_group

Gets the full metadata of one business group from its identifier.

When it helps
a new colleague asks who owns the EMEA business group and which environments it contains, and nobody wants to dig through Access Management for that on a busy morning.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

get_mcp_server_state

Reads the deployed state of a catalog-managed tool server: its current systems, tools and upstream identifiers, the input that any later adjustment depends on.

When it helps
before adding a system to an existing server, Claude checks what is running today, so nothing already live is overwritten by mistake.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

get_omni_gateway_target_domains

Gets the ingress domains and application identifier of one managed gateway target.

When it helps
the network team needs the exact domains of a new gateway target to set up their firewall rules before go-live, instead of waiting on a reply from the platform team.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

get_omni_gateway_usage_report

Reports how much of your managed gateway size entitlement is currently in use.

When it helps
before launching another gateway, you want to know whether the contract still has room or whether procurement must be looped in first.
Watch out
the report covers size entitlement usage, not traffic or cost.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

get_policy_template_form

Loads one policy template's details, including its configuration schema, so the matching form can be shown right after you pick a policy.

When it helps
you chose rate limiting and want to see which fields, limits and time windows it will ask for.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

get_provision_status

Polls the progress of a provisioning job, or of a job that adjusts a tool server.

When it helps
Claude started deploying a server a minute ago and you want to know whether it is live before telling the product team it is ready.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

get_trusted_mcp_catalog

Returns the catalogue of SaaS systems available for tool server provisioning, each with its tools, authentication type, credential fields and upstream settings.

When it helps
you want to know whether Salesforce or ServiceNow can be wired into a composite server before planning it.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

list_agents

Lists the agents registered for your organization and presents a browsable view with filters for picking one.

When it helps
a compliance officer wants every registered agent in one place before an AI inventory is sent to the regulator, including agents that business teams registered themselves.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

list_apis

Lists the APIs in your organization's catalogue with an interactive view for browsing, filtering and selecting them.

When it helps
an architect wants to see every API tagged for the finance domain before drawing the new integration map for the next quarter.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

list_business_groups

Lists the business groups available to you, optionally filtered by name.

When it helps
your organization has dozens of groups and you need the right one before any other question makes sense.
Watch out
listing does not switch anything; select_active_business_group is the tool that sets the active group.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

list_governance_strategies

Lists the governance strategies of your organization in a searchable, filterable table styled like the main web app.

When it helps
before proposing a new strategy, you check whether one already covers PII detection for customer-facing agents, which would make a second one redundant.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

list_llms

Lists the language models registered for your organization, with a view to browse, filter and select them.

When it helps
the data team asks which models are approved across business units before a new chatbot project starts, so nobody picks an unapproved one by habit.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

list_mcp_servers

Lists the tool servers registered for your organization, the ones that give AI agents access to systems, with a browsable view.

When it helps
security wants an inventory of every server agents can call before the annual review, including servers that were set up for a single pilot.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

list_omni_gateway_environments

Lists the Anypoint environments available for managed gateway setup, so the right one can be chosen first.

When it helps
you plan a gateway for staging and want to confirm the environment exists and is eligible before anyone opens the wizard.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

list_omni_gateway_targets

Lists the deployment targets available for managed gateways in one environment.

When it helps
a gateway must run in a specific region for data residency, and you want to see which targets are offered there before committing to a deployment plan.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

list_provider_gateways

Lists gateways for the active provider; with MuleSoft as provider, that means your Flex Gateways.

When it helps
you want to know how many Flex Gateways run in production and where, without opening Runtime Manager for each environment one after the other.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

list_scanner_providers

Gets the list of providers a scanner can connect to, such as Google Vertex or Microsoft Copilot.

When it helps
a team building agents on another platform wants to know whether MuleSoft can discover them automatically before planning a manual inventory.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

prepare_automated_policy_creation

Opens the wizard for a new automated-policy strategy, with the scope options and policy templates to choose from.

When it helps
you want to see which scopes and templates exist before deciding whether an automated approach fits your estate at all.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

prepare_policy_creation

Loads the policy options available for one API instance, the starting point for adding protection to it.

When it helps
asked to secure a Stripe-facing API, Claude first shows which policies that instance accepts, so you choose from real options rather than from memory.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

search_assets_semantic

Searches portfolio assets by meaning, for concepts or use cases rather than exact words, and returns names, descriptions, relevance scores and types.

When it helps
you ask for services related to customer authentication and get candidates even when none is named that way.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

search_global_content

Searches content across agents, tool servers and APIs of your platform in one go, a tool MuleSoft designed for embedded agents.

When it helps
you remember a component was called something like order sync but not whether it is an API or an agent.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

search_portfolio_services

Discovers assets from the capabilities and use cases they cover, a way to explore the portfolio by need.

When it helps
a project manager needs anything that can check an address or a VAT number and has no idea what those services are called internally.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

search_repository_knowledge

Answers conceptual questions from a curated, fixed corpus of policy, governance and security documents and rulesets, with cited snippets.

When it helps
someone asks what a given governance ruleset checks, and you want the cited passage instead of a paraphrase.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

select_active_business_group

Shows the current business group context and, when given an identifier, sets a new active group for the tools that follow. No source says it touches any Anypoint resource.

When it helps
switching from the retail group to the logistics group mid-conversation.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

select_active_environment

Browses the Anypoint environments of the active provider and sets one as the working context for later calls.

When it helps
you want every following question answered against production, not sandbox, without repeating it in each request you make to Claude.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

select_active_provider

Browses the available providers and sets one as the active context for the provider tools that come next.

When it helps
your gateways span MuleSoft and another provider, and the next questions concern only the MuleSoft side of the estate.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

select_api_version

Shows an interactive selector of versions for one API and lets the host know which version you pick.

When it helps
an API exists in versions one, two and three, and the question only makes sense for the one partners still call.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

show_observability_performance

Loads observability dashboard data so it can be displayed as an embedded view inside the conversation.

When it helps
during an incident call, you want the performance dashboard in front of everyone without anyone sharing a screen from the console.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

test_scanner_connection

Tests the connection to a scanner provider before the scanner itself is set up, so a wrong credential shows up early.

When it helps
the cloud team handed over new Vertex credentials and you want to know they work before running the wizard.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

view_agent_details

Shows one agent's details in read-only form: versions, instances and metadata.

When it helps
the invoice-approval agent behaves differently in two regions, and you want to compare which version each instance runs before opening a ticket with its owners in the business team.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

view_api_instance_details

Displays one API instance with a header matching the web app, from the instance itself or from its identifiers.

When it helps
you need the endpoint, environment and status of the instance a partner reports as failing, before calling them back.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

view_api_instance_monitoring

Presents monitoring charts, traffic and clients for one API instance, as a visual panel rather than raw figures.

When it helps
a product owner wants to see who actually calls their API and how much, ahead of a deprecation decision.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

view_api_instance_policies

Displays the policies applied to one API instance, grouped the same way as in the web app.

When it helps
before adding rate limiting, you check whether a similar policy is already in place on that instance so two rules do not collide.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

view_api_version_details

Shows one API version in read-only form, with its instances, metadata and lifecycle fields.

When it helps
you want to know whether version two of the customer API is still active or already marked for retirement before advising a partner team.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

view_api_version_governance_report

Presents the governance ruleset report summary and its findings for one API version.

When it helps
a release is blocked on governance and the team wants the list of findings in plain words, not a scorecard, before the go or no-go meeting.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

view_api_version_instances

Lists the instances of one API version and lets you select some of them to hand back to the conversation for the next step.

When it helps
a policy should go on two of the four instances only, and you want to pick them visually.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

view_governance_report

Fetches the latest governance strategy report for your organization and displays it in an interactive panel inside the chat.

When it helps
the monthly risk committee wants the current governance picture without anyone exporting spreadsheets from the console beforehand, and questions get answered on the spot.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

view_llm_details

Shows one language model's details in read-only form: versions, instances and metadata.

When it helps
legal asks which version of a model runs behind the HR assistant and where it is deployed, ahead of a data protection review that needs precise answers.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

view_mcp_server_details

Shows one tool server's details in read-only form, with its versions, instances and metadata.

When it helps
before connecting a new agent to a server, the architect checks which version is deployed in each environment, to avoid pointing the agent at a test build.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

What Claude changes (9)

9 tools

Nine tools that apply policies, set up gateways, scanners and servers, or delete strategies. No source describes a Claude confirmation specific to them.

apply_policy_to_instance

Approval: see the rule

Puts the selected policy in place on the target API instance, using the configuration you filled in.

What Claude asks for
no source describes a confirmation for this tool, so the general approval rule further down applies.
When it helps
rate limiting is needed on a partner-facing API before a busy launch week.
Watch out
in MuleSoft's own example, the policy takes effect immediately.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

create_automated_policy_strategy

Approval: see the rule

Sets up a governance strategy that applies policies automatically, from the answers given in its wizard.

What Claude asks for
nothing tool-specific is documented; see the approvals section for the rule that applies.
When it helps
every new customer-facing API should get the same baseline protections without anyone remembering to add them.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

create_omni_gateway

Approval: see the rule

Opens the managed gateway wizard for the active MuleSoft provider and, once fully configured, sets up the gateway in Anypoint Platform.

What Claude asks for
no source says whether a prompt shows before this call; the default rule in the approvals block covers it.
When it helps
a new environment needs its own gateway before APIs can be exposed there.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

create_scanner

Approval: see the rule

Runs the scanner wizard, which walks through provider choice, authentication and configuration, then returns the new scanner's details.

What Claude asks for
the sources stay silent on a confirmation here, so lean on the general approval rule.
When it helps
agents built on Google Vertex or Microsoft Copilot should show up in your catalogue without manual entry.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

delete_automated_policy_strategy

Approval: see the rule

Removes one automated-policy governance strategy, picked by its identifier.

What Claude asks for
none of the sources describe a prompt for this deletion; treat the approvals block as your safety net.
When it helps
a pilot strategy tested on a sandbox group is no longer wanted.
Watch out
no source mentions a way to restore it.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

delete_governance_strategy

Approval: see the rule

Takes one governance strategy away, identified by its strategy identifier.

What Claude asks for
as with the other deletion, no confirmation specific to it is documented, and only the default rule applies.
When it helps
two overlapping strategies flag the same APIs and one of them has to go.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

provision_mcp_server

Approval: see the rule

Deploys a composite tool server on Flex Gateway, assembled from systems picked in the trusted catalogue.

What Claude asks for
no source describes a Claude-side prompt for this deployment; the general rule covers it.
When it helps
agents need governed access to Salesforce and Jira through one endpoint.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

resume_provision_job

Approval: see the rule

Picks a failed provisioning or update job back up from its last checkpoint, provided some steps had already completed.

What Claude asks for
the Claude-side prompt is undocumented for this call as well; rely on the approvals block.
When it helps
a deployment stopped halfway on a timeout and you do not want to start over.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

update_mcp_server

Approval: see the rule

Brings an existing catalog-managed server in line with a new configuration by comparing it with what is deployed and applying only the differences; it returns a job to follow.

What Claude asks for
no confirmation specific to this tool is described, so the general rule applies.
When it helps
adding one system to a live server.
Watch out
it needs the current deployed state first.

Sourcedocs.mulesoft.com · September 28, 2026 ↗

Not documented (5)

5 tools

The directory lists these five names, but MuleSoft's reference dated September 2026 does not. This page will not guess what they do.

check_control_scope_targets

This name sits in the directory's tool list and nowhere in MuleSoft's current reference. No official source describes it, so this page says nothing more about it and makes no guess.

Watch out
whatever it does, the general approval rule covers it.

create_control

Listed on the directory sheet, absent from MuleSoft's tool reference: no documentation describes this one either. Filling the gap with a guess from its name would be misleading.

Watch out
it falls under the same approval rule as the rest of the list.

fetch_control_ruleset

Here again the directory shows a name that MuleSoft's reference does not contain. With no official description available, the note stops at that fact and draws no conclusion from the name.

Watch out
the approvals section below applies to it all the same.

prepare_control_creation

One more name from the directory with no entry in MuleSoft's reference and no description anywhere official. The page leaves it undocumented rather than inventing a role for it on the strength of a name.

Watch out
the general approval rule applies to it like the others.

view_service_governance_findings

The fifth name the directory lists without any match in MuleSoft's tool reference. No official text describes it, so this page does not either, and no role is assumed from the wording.

Watch out
the general approval rule covers it too.
Approvals

What Claude asks you before acting

By default, Claude stops and asks for your go-ahead before each action it takes on an account for you. No source describes a MuleSoft-specific confirmation, so that rule covers policies, deployments and deletions.

Anypoint adds its own gate. The connected app acts on your behalf, and MuleSoft says some tools need additional permission scopes, tied to Anypoint roles such as API Manager Environment Admin or Governance Administrator. Without the right one, a call cannot go through, whatever Claude does. On Team or Enterprise, Claude's owners decide whether members may skip prompts and can restrict the connector for everyone.

Plans

Which plans this is available on

Out of 819 directory sheets, not one shows availability by plan. The connector-by-connector answer is published nowhere: it is a real gap in the catalogue.

The general rule is public: remote connectors are open to all users on Claude, Cowork, Claude Desktop and mobile. On Team and Enterprise, an Owner or Primary Owner turns the connector on for the organization before members can connect. On the MuleSoft side, the connected app and Anypoint permissions set what works. For this connector's current state, check its sheet in the official directory.

Limits

Where this connector stops

A connector is not an automation. Claude calls these tools while it answers you: no scan, report or policy runs because a new API was published.

Everything known here comes from MuleSoft's documentation and the directory sheet; Claude's help centre has no page on this connector. The tool list is an observed floor, not a promise, and MuleSoft's reference already mentions tools the sheet does not show. The verification badge is not a security audit either: Anthropic states on every sheet that it does not choose the tools a publisher exposes. The Integrations hub gathers the other ways to connect MuleSoft.

Need help

Need help connecting MuleSoft to Claude?

A person reads every message.

FAQ

Frequently asked questions about the Claude MuleSoft connector

01What can Claude do with the MuleSoft connector?
Claude can explore and govern your Anypoint landscape from a conversation. The connector lists sixty-six tools: searching and browsing APIs, agents, language models and tool servers, reading monitoring, cost and governance reports, preparing and applying policies, setting up gateways and agent scanners, deploying or updating composite tool servers, and deleting governance strategies. Fifty-two tools only read, nine act on the platform, and five appear on the directory sheet without any description in MuleSoft's reference.
02Can Claude apply policies or delete things in MuleSoft?
Yes, nine tools act on your platform. They apply a policy to an API instance, set up an automated-policy strategy, launch a managed gateway, configure an agent scanner, deploy, resume or update a composite tool server, and delete an automated-policy strategy or a governance strategy. MuleSoft's example shows a policy taking effect immediately. MuleSoft adds that some tools need extra Anypoint permission scopes, so a user without them cannot go through Claude either.
03Does Claude ask before it applies a policy?
By default, yes: Claude asks for confirmation before each action it takes on an account for you. No source describes a confirmation specific to a MuleSoft tool, so that default rule covers policies, deployments and deletions. Since MuleSoft says a policy applies at once, read the prompt carefully. On a Team or Enterprise workspace, Claude's owners decide whether members may let actions through without a prompt and can close writing entirely.
04Which Claude plans is the MuleSoft connector available on?
No official source publishes plan availability connector by connector, and no directory sheet shows it. The general rule says remote connectors are open to all users on Claude, Cowork, Claude Desktop and mobile. On Team and Enterprise, an Owner or Primary Owner has to enable the connector before members can sign in individually. On MuleSoft's side, what works depends on your connected app and Anypoint permissions. The connector's sheet in the official directory shows its current state.
05Does Claude see everything in our Anypoint organization?
No. The connected app acts on behalf of the person who signs in, so Claude sees what that person's Anypoint permissions allow. MuleSoft says some tools need additional permission scopes, with Anypoint roles such as Exchange Viewer or Monitoring Viewer. Within those rights, the active business group and environment are set with dedicated selection tools. The connected app itself needs the Full Access and Background Access scopes. Team or Enterprise owners can further restrict the connector in Claude.
06Why won't the MuleSoft connector connect to Claude?
Most often because the connected app is missing or incomplete. MuleSoft only supports manual client registration: someone must create a connected app in Anypoint Platform with the Full Access and Background Access scopes, and Claude's callback address among its redirect URLs. You then enter its Client ID and Client Secret when adding the connector. Automatic client registration is not supported, so skipping this step blocks the sign-in. The callback address goes in the connected app's Redirect URLs field.
07Claude or an automation tool for MuleSoft?
They answer different needs, so it depends on the task. Claude works inside a conversation: you ask about your APIs, costs or governance, it looks them up and can act when you agree, and nothing happens once you stop typing. It does not react on its own when an API is published or an error rate climbs. For scheduled or event-driven CloudHub jobs, that is an automation tool's role. For investigating and governing on demand, the connector fits.