Bragproof

← Reports

@doitintl/doit-mcp-server

Starts with no setup5.0.0npm

Installed, started with no arguments, no environment and no input, and listed its tools. Run 2026-11-16.

Needs to start: nothing
Returns a result when called with no setup: 0 of 5 tools tried
Contacts while installing and starting: nothing

200tools listed
0hosts contacted
0files written
0.3 sto start
4dependencies

Its tool definitions changed since version 4.0.0. If you approved the earlier ones, read them again.

What it contacted

Nothing besides the npm registry it was installed from. Watched from outside its sandbox while it was installed and started with no input. Its tools may still contact things when called.

Starting it

npx -y @doitintl/doit-mcp-server@5.0.0

Installing a package runs its code. Use a sandbox for anything you have not read.

Calling its tools

5 tools that take no input and look read-only were called once with nothing. 0 returned a result.

list_cloud_connectionsNo usable answer tool reported an error: Unauthorized
get_cloud_overviewNo usable answer tool reported an error: Unauthorized
get_cloud_incidentsNo usable answer tool reported an error: Unauthorized
get_cloud_incidentNo usable answer tool reported an error: Unauthorized
get_anomaliesNo usable answer tool reported an error: Unauthorized

During the calls it contacted nothing it had not already contacted.

Only tools whose names read like looking something up, with no required input, are tried, five at most. Tools that need input, and anything that sounds like it changes or sends something, are not called.

What its tools tell an agent

In the tool descriptions, nothing that talks about the user, other tools or private files, or uses hidden characters.

3 tools with commands in capitals or very long text
  • Very long description in run_query: 2759 characters or more
  • Very long description in list_customer_subscriptions: 2628 characters or more
  • Very long description in update_customer_subscription: 3000 characters or more

Common and usually harmless: it is how many makers stress how a tool should be used.

Found by simple pattern rules in each tool's description (up to 3,000 characters), quoted as written. Parameter descriptions are not scanned. A clean result is not a safety review.

Tools

get_cloud_connect_requirements scope, featureIds, featureInputs, identityInputs, discoveryInputs
Resolve fresh customer-specific AWS, GCP or Azure Cloud Connect prerequisites, identity trust, per-feature permissions, inputs and deployment artifacts. Read-only. Azure management-group requirements return an inline Azure CLI plan/apply script with exact selected roles and a customer binding; it grants access without registering and does not elevate the operator. GCP organization, project and dataset scopes use the returned inline shell artifact. AWS account scopes return the account CloudFormation template and per-account Terraform module. AWS organization scopes return one management StackSet artifact with entrypoint, a flat files list (path, URL and SHA-256 for both templates) and parameters for one root/OU; no fleet connection or rollout operation is created. AWS membership, Organizations/StackSets prerequisites, deployment success and access remain separate checks. AWS organization real-time bucket/enrollment configuration is unsupported by the pinned artifact. Artifacts are reference implementations for adaptation to customer tooling; their requirements and registration behavior remain relevant when adapted. The selected artifact's requirementsSatisfied and findings determine artifact readiness; feature-level flags or preconditions alone do not. Discovery and blocked artifacts have requirementsSatisfied=false and empty parameters. Returned parameters preserve the artifact's exact spelling and structure. Does not enable billing exports. Full Cloud Connect workflow: MCP resource skill://doit-cloud-connect/SKILL.md.
list_cloud_connections provider, scope, featureId, maxResults, pageToken
Discover existing Cloud Connect records by structured provider, scope and featureId filters. Pass provider azure or gcp, or that provider’s exact scope, for non-AWS discovery; omitting both provider and scope defaults to AWS. An empty list means no matching registration, not that cloud resources are absent. Subsequent pages use pageToken with the same filters.
get_cloud_connection connectionId
Read a connection's bindings, desired features, last verification summary and named follow-ups. Returns the etag for conditional writes. Stored state is not a fresh access probe; fresh checks are available through verify_cloud_connect.
register_cloud_connection identity, bindings, catalogVersion, idempotencyKey
Register a deployed AWS IAM role, GCP service account or Azure DoiT enterprise application explicitly when the final deployment does not register it automatically and no connection exists yet. For AWS, retained DoiT Terraform provider resources and CloudFormation callbacks perform registration through their own workflow; an additional registration duplicates that operation. An existing connection for the scope returns 409 connection_exists; changes to that record use a conditional update. Fresh requirements describe prerequisites, and dry_run_cloud_connection_change previews the request. GCP setup scripts grant access but do not register: use descriptor metadata and the selected GCP scope. Azure uses identity.kind azure_service_principal with tenantId and one management-group binding; the live selected grants must carry the setup-written customer binding. The response returns the authoritative connection id, bindingId and registered scope identifiers for subsequent operations. Registration records an existing cloud target; it does not create an account, project, organization or Azure tenant/group. Registration alone does not establish feature readiness or data arrival. Full Cloud Connect workflow: MCP resource skill://doit-cloud-connect/SKILL.md.
update_cloud_connection connectionId, expectedEtag, catalogVersion, bindings, idempotencyKey
Change a binding's desired features/inputs, or retry failed follow-ups by re-sending the current desired state. Requires the latest connection etag; 412 indicates a stale value and a change that needs re-evaluation. dry_run_cloud_connection_change previews the exact request and its effects, including removals. Does not change IAM or the identity. Full Cloud Connect workflow: MCP resource skill://doit-cloud-connect/SKILL.md.
195 more tools
remove_cloud_connection connectionId, expectedEtag, cleanupScope, idempotencyKey
Disconnect the DoiT record using its exact etag and cleanupScope doit_only; customer IAM/IaC cleanup remains separate. dry_run_cloud_connection_change previews removal effects. 204 means the record is gone; a 409 cleanup_failed keeps it, and another removal attempt requires a fresh etag. A later 404 means already absent. Full Cloud Connect workflow: MCP resource skill://doit-cloud-connect/SKILL.md.
dry_run_cloud_connection_change
Validate a register, update or remove request and run available live checks without committing (dryRun=true). A new, unregistered AWS role returns aws_broker_requires_registration because the broker cannot probe it before registration; this result does not verify access. Returns preconditions, DoiT side effects, follow-ups and customer cleanup for review. wouldApply is informational, not a reservation or approval. Full Cloud Connect workflow: MCP resource skill://doit-cloud-connect/SKILL.md.
verify_cloud_connect connectionId, bindingIds, featureIds
Run fresh synchronous access checks for a registered connection. Results include authentication, per-feature checks and coverage, and `recorded` indicating whether the observation was stored. partial, not_checked and failed checks are not passed; a failed authentication leaves every feature not_checked. AWS declared-policy checks do not prove effective access. GCP checks effective scope permissions; BigQuery dataset results expose sampled table coverage. Azure checks fresh management-group access and unconditional role assignments held by the DoiT application at that exact scope; this does not evaluate every effective permission or deny assignment. Access never proves billing data arrived. Nothing is stored as a run; requestId provides correlation. Full Cloud Connect workflow: MCP resource skill://doit-cloud-connect/SKILL.md.
get_cloud_overview
Use this when the user wants a high-level overview or dashboard of their entire cloud infrastructure. Returns cost by cloud provider, top services per cloud, top projects per cloud, recent cost anomalies, and recent cloud incidents — all in a single call. Costs cover the last 30 days including today as totals per cloud, per service, and per project (no daily rows); services and projects are limited to the top 5 values per cloud across that range. Returns at most 5 anomalies and 5 AWS/Google Cloud provider incidents (not account-specific incidents). Failed sections silently return empty arrays, so empty data does not establish zero spend or no incidents. Do NOT use this for detailed drill-downs (use run_query), single-provider analysis, or anomaly-only queries.
get_cloud_incidents platform, filter, pageToken
Use this when the user wants to check for active cloud platform outages, service disruptions, or incidents from AWS or Google Cloud. Both active and archived incidents are included unless status is filtered. Do NOT use this for cost anomalies (use get_anomalies) or support tickets (use list_tickets).
get_cloud_incident id, title
Use this when the user wants to view details of a specific cloud platform incident. Accepts either the incident ID or a partial title match (case-insensitive). Do NOT use this for listing all incidents (use get_cloud_incidents) or anomalies (use get_anomalies).
get_anomalies pageToken
Use this when the user wants to check for unexpected cost spikes, billing anomalies, or unusual spending patterns. Returns recent anomalies with severity and impact. Do NOT use this for optimization recommendations or savings opportunities (use list_optimization_recommendations), regular cost analysis (use run_query), or viewing alerts (use list_alerts).
get_anomaly id
Use this when the user wants to view details of a specific cost anomaly by its ID. Returns full anomaly data including affected resources and cost impact. Do NOT use this for listing all anomalies (use get_anomalies).
list_reports filter, pageToken, name
Use this when the user wants to see their saved Cloud Analytics reports or browse available reports. Returns pages of up to 40 reports with IDs and metadata, newest creation first. rowCount counts this page; pageToken retrieves the next page and is null at the end. Empty results return an error. Supports server-side, case-insensitive partial name filtering through name. Do NOT use this for running queries (use run_query) or getting report results (use get_report_results).
run_query config
Use this when the user wants to analyze cloud costs, generate a cost breakdown, view spending trends, or run a custom analytics query across their cloud providers. Runs the config through the DoiT Cloud Analytics API query endpoint (https://developer.doit.com/reference/query) and returns the result rows. Accepts a structured config with data source, metrics, dimensions, time range, and filters. Do NOT use this for listing saved reports (use list_reports), checking anomalies (use get_anomalies), or viewing budgets (use list_budgets).
    Unpopulated fields take API defaults: basic cost, last 7 days including today, daily time rows (year/month/day), and billing (or billing-datahub for customers with DataHub metrics).
    Each `config.group[].limit.value` selects top/bottom N dimension values per parent group, ranked across the range; it does not cap result rows and has no tool-enforced maximum of 25. Time columns can produce multiple rows per group.
    `timeRange` covers relative periods; explicit dates require `timeRange: {mode: "custom"}` and sibling `config.customTimeRange: {from, to}`, with no unit.
    For mode "last", "includeCurrent": true includes the current partial period within amount; false selects fully completed periods. One month with true is current month-to-date.
    "metrics" (array) supersedes the deprecated "metric" (object).

    A "group" with id "service_description" and type "fixed" returns a per-service cost breakdown, the most common shape for cost questions.

    Common grouping dimension IDs (all type "fixed"):
      "service_description" — cloud service
      "project_id"          — GCP project / AWS account / Azure subscription
      "cloud_provider"      — cloud provider (AWS / GCP / Azure)

    In default mode "is", filter values must exactly match stored dimension values (case-sensitive): cloud_provider uses provider IDs, while service_description uses service names. Common cloud-provider aliases such as aws/gcp/azure are normalized by run_query. get_dimension({type, id}) returns the valid values for a dimension for this customer.
    Known cloud provider IDs (cloud_provider, type "fixed"):
      "amazon-web-services" = AWS, "google-cloud" = GCP, "microsoft-azure" = Azure

    Example — top AWS services last month:
    {
      "config": {
        "dataSource": "billing",
        "metrics": [{"type": "basic", "value": "cost"}],
        "timeRange": {"mode": "last", "amount": 1, "unit": "month", "includeCurrent": false},
        "filters": [{"id": "cloud_provider", "type": "fixed", "values": ["amazon-web-services"]}],
        "group": [{"id": "service_description", "type": "fixed", "limit": {"metric": {"type": "basic", "value": "cost"}, "sort": "desc", "value": 10}}]
      }
    }
cost_breakdown groupBy, cloud, months, topN
Use this when the user wants a simple cost breakdown by service, project, or cloud provider (e.g. 'What are my top services by cost?', 'Which projects cost the most?'). Selects the top-N groups by total cost across the range and returns monthly rows for each. The range includes the partial current month; months=1 means month-to-date. Output row order is not a cost ranking. For complex multi-filter or multi-metric queries, use run_query instead.
cost_trend months, cloud, groupBy, topN
Use this when the user wants to see monthly spend over time (e.g. 'Show me my cost trend', 'How has my spend changed over the last 6 months?'). Returns monthly cost data points, optionally broken down by service/project/cloud. The last point is the partial current month; months=1 means month-to-date. For daily granularity or custom time intervals, use run_query instead.
compare_spend period1Months, period2, cloud, groupBy
Use this when the user wants to compare spend between two time periods (e.g. 'Compare the latest three months including this month with January through March'). Period 1 is N-1 full calendar months plus current month-to-date; period 2 is an explicit date range. Returns two separate sets of monthly rows, each independently limited to its top 10 groups across that period. Groups can differ; no difference or percentage change is computed. For more than two periods or advanced comparative analysis, use run_query instead.
list_optimization_recommendations category, priority, displayStatus, provider, easyWin, searchTerm, page, pageToken, maxResults, pageSize
Use this when the user asks about optimization, recommendations, insights, savings opportunities, rightsizing, idle resources, security findings, or cost reduction suggestions. Also use this when the user asks 'what insights are available?' or 'show me insights'. This is the primary tool for 'what can I optimize?', 'how can I save money?', and 'what insights do I have?' questions. Returns insights ordered within each page by estimated daily savings, not priority. All statuses are included unless filtered, including dismissed. easyWin is derived from easyWinDescription when present, otherwise null. Follow pageToken for more results. Do NOT use this for cost anomalies/spikes (use get_anomalies) or budget tracking (use list_budgets).
get_insight_resources pageToken, maxResults, source, key
Use this when the user wants to see which specific resources are affected by an optimization insight. Returns a page of resourceResults with rowCount and pageToken, resource IDs, accounts, savings or risk counts. There is no dedicated remediation field. Default page size is 1000; follow pageToken for more resources. Do NOT use this for listing all insights (use list_optimization_recommendations).
get_insight source, key
Use this when the user wants the details and aggregate summary (savings, risk counts, status, description) of a single optimization insight identified by its source and key. Returns the insight metadata only — it does NOT include the individual affected resources (use get_insight_resources for those) and is not for listing all insights (use list_optimization_recommendations).
post_insight_result source, key, title, shortDescription, cloudProvider, categories, detailedDescriptionMdx, reportUrl, cloudFlowTemplateId, easyWinDescription, status, dismissalDetails
Use this when the user wants to create a new custom insight or update an existing one's metadata (title, description, categories, status, remediation links). Only insights owned by the 'public-api' source can be managed (the default source). Requires key, title, shortDescription, cloudProvider and categories on every call. This replaces insight metadata: omitted optional fields such as detailedDescriptionMdx, reportUrl, easyWinDescription and cloudFlowTemplateId are cleared. Affected resources are managed separately (post_insight_resource_results). For a status-only change, use update_insight_status to preserve metadata despite that endpoint's deprecation.
update_insight_status source, key, status, dismissalDetails
Use this when the user wants to change the display status of an existing insight (e.g. mark it acknowledged, in progress, optimized, or dismissed). This endpoint is deprecated but remains the safe choice for status-only changes because it preserves optional insight metadata; post_insight_result replaces it. Only insights owned by the 'public-api' source can be managed. When dismissing, dismissalDetails.reason is required; comment is optional. Do NOT use this to edit an insight's title/description or create one (use post_insight_result).
get_report_results id, name
Use this when the user wants to retrieve the data results of a specific saved report. Uses the report's saved time range, resolved at execution for relative ranges. Accepts an ID or a case-insensitive substring name lookup within the first 200 newest reports; multiple matches return an ambiguity error and id takes precedence. Do NOT use this for listing all reports (use list_reports) or running ad-hoc queries (use run_query).
get_report_config id
Get the configuration of a specific Cloud Analytics report by ID. Returns the stored report object including name, type, and a nested 'config' field containing data source, metrics, dimensions, time range, filters, and visualization settings.
create_report name, description, labels, folderId, config
Use this when the user wants to save a new Cloud Analytics report with a specific configuration. Creates a new custom report immediately without replacing existing reports, optionally inside a folder (folderId). Omitted config fields take API defaults: basic cost, last 7 days including today, daily time rows, and billing (or billing-datahub when the customer has DataHub metrics). System labels are rejected. Do NOT use this for one-time queries without saving (use run_query).
update_report id, name, description, labels, folderId, config
Use this when the user wants to modify an existing saved Cloud Analytics report. Only custom reports the caller can edit are supported. Supports partial updates with array replacement and config reset semantics described on config. Omitted top-level fields are unchanged; labels replaces non-system labels; folderId moves the report to another folder. Changes apply immediately. Do NOT use this for running ad-hoc queries (use run_query).
validate_user
Use this when the user asks to verify their account connection or check who they are logged in as. Returns the authenticated user's email and the primary domain of the customer the session is scoped to. Authentication is already established by the connection, so other tools do not depend on this call. Do NOT use this for listing users in the organization (use list_users).
list_dimensions filter, pageToken
Use this when the user wants to see available dimensions for cost analysis queries. Returns id, label and type metadata only (no values), up to 200 dimensions per page, sorted by id. GKE dimensions are omitted. pageToken retrieves subsequent pages; rowCount counts this page. get_dimension retrieves values. Empty results return an error. Do NOT use this for running cost queries directly (use run_query) or viewing allocations (use list_allocations).
get_dimension type, id
Use this when the valid filter values for a specific dimension are needed, such as for a run_query filter, or when the user wants to view dimension details. Returns id, label, type and all available value/cloud pairs in one response, without pagination. For example, get_dimension({type: 'fixed', id: 'cloud_provider'}) returns the exact provider IDs available for this customer. Do NOT use this for listing all dimensions (use list_dimensions) or running queries (use run_query).
list_tickets pageToken, pageSize, subject
Use this when the user wants to view their support tickets, check ticket status, or review open issues. Returns tickets with status, severity, and platform. Customers see their own tickets; organization tickets are visible when ticket sharing is enabled. Subject filtering is case-insensitive and applies only to the returned page. The API cursor and unfiltered rowCount are preserved, including on pages with no matches. Each ticket's fields array lists only custom fields that have a value; unset fields are omitted. Do NOT use this for cloud incidents (use get_cloud_incidents) or cost alerts (use list_alerts).
get_ticket id
Returns details of a specific support ticket from the DoiT API by its ID.
list_ticket_comments ticketId, pageToken
Returns all comments on a support ticket. For customers, only public comments are returned. For DoiT employees, both public and private comments are returned.
create_ticket_comment ticketId, body, private
Adds a comment to an existing support ticket. For customers, comments are always public. For DoiT employees, comments can be marked as private (internal notes) by setting the private field to true.
create_ticket ticket
Use this when the user wants to create a new support ticket. The ticket is opened with DoiT support immediately. Do NOT use this for viewing existing tickets (use list_tickets) or cloud incidents (use get_cloud_incidents).
list_invoices pageToken
Use this when the user wants to see their invoices, check billing history, or review payment records. Returns a list of invoices with amounts, dates, and status. Do NOT use this for cost analysis (use run_query) or budget tracking (use list_budgets).
get_invoice id
Use this when the user wants to view details of a specific invoice by its ID. Returns full invoice data including line items and status. Do NOT use this for listing all invoices (use list_invoices) or cost analysis (use run_query).
list_allocations pageToken, name
Use this when the user wants to see their cost allocation rules or configurations. Returns a list of allocations. Returns pages of 40; pass the returned pageToken for another page (null means no next page). The case-insensitive partial name filter applies only to that returned page. Do NOT use this for cost queries (use run_query) or labels (use list_labels).
get_allocation id, name
Use this when the user wants to view details of a specific cost allocation. Accepts either the allocation ID or a case-insensitive partial name. Name lookup searches only the first 200 allocations; multiple matches return an error listing names, and id takes precedence. Do NOT use this for listing all allocations (use list_allocations) or running queries (use run_query).
create_allocation name, description, rule, rules, unallocatedCosts
Use this when the user wants to create a new cost allocation rule. Changes apply immediately. Do NOT use this for viewing existing allocations (use list_allocations) or labels (use create_label).
update_allocation name, description, rule, rules, unallocatedCosts, id
Use this when the user wants to modify an existing cost allocation. Omitted fields are preserved. Use rule for an existing single allocation and rules for an existing group; the type cannot change. A supplied rules list replaces the entire group membership. Changes apply immediately. Do NOT use this for creating new allocations (use create_allocation) or viewing allocations (use list_allocations).
list_assets maxResults, pageToken, filter, name
Use this when the user wants to browse their cloud assets, subscriptions, or resources. Returns a paginated list of assets. Name filtering is case-insensitive and applies only to the returned page. The API cursor and unfiltered rowCount are preserved, including on pages with no matches. Do NOT use this for cost analysis (use run_query) or checking invoices (use list_invoices).
get_asset id, name
Use this when the user wants to view details of a specific cloud asset. Accepts an asset ID, which takes precedence over name, or a case-insensitive partial name lookup within the first 249 assets only. Multiple name matches return an ambiguity error. Do NOT use this for listing all assets (use list_assets) or cost analysis (use run_query).
list_alerts sortBy, sortOrder, maxResults, pageToken, filter
Use this when the user wants to see their cost alerts or check alert configurations. Returns a paginated list of alerts. Do NOT use this for anomaly detection (use get_anomalies) or budget tracking (use list_budgets).
get_alert id, name
Use this when the user wants to view the details of a specific cost alert. Accepts either the alert ID or a partial name (case-insensitive). Do NOT use this for listing all alerts (use list_alerts) or anomalies (use get_anomalies).
create_alert name, config, recipients
Use this when the user wants to set up a new cost alert with thresholds and notification settings. Changes apply immediately. Do NOT use this for creating budgets (use create_budget) or viewing existing alerts (use list_alerts).
update_alert id, config, name, recipients
Use this when the user wants to modify an existing cost alert. Supports partial updates, including name-only, recipients-only, or individual config fields. Omitted fields are preserved; scopes: [] clears scopes. Changes apply immediately. Do NOT use this for creating new alerts (use create_alert) or budgets (use create_budget).
trigger_cloud_flow flowID, requestBodyJson
Triggers a published CloudFlow with a webhook trigger by flow ID or trigger URL. Executes real actions immediately and returns executionLink. Drafts fail with 403, flows without a webhook trigger with 400, and already-running flows with 409. Manual or scheduled flows use trigger_cloudflow_flow instead. This may modify external cloud resources.
list_cloudflows maxResults, pageToken
Use this when the user wants to see their CloudFlow automation flows. Returns a cursor-paginated list of flows with their metadata, status, and last execution info.
list_cloudflow_connections maxResults, pageToken
Use this when the user wants to see their CloudFlow cloud provider connections (the GCP/AWS accounts connected for automation). Returns a cursor-paginated list of connections with their config and status. Do NOT use this to trigger a flow (use trigger_cloud_flow) or to view a single connection's details (use get_cloudflow_connection).
get_cloudflow_connection connectionId
Use this when the user wants to view the details of a specific CloudFlow cloud provider connection by its ID, including its GCP/AWS configuration, collaborators, and status. Do NOT use this to list all connections (use list_cloudflow_connections) or to trigger a flow (use trigger_cloud_flow).
create_cloudflow_connection idempotencyKey, name, description, gcpConfig, awsConfig, collaborators, enabled
Use this when the user wants to create a new CloudFlow cloud provider connection (a GCP or AWS account connected for automation). Exactly one of gcpConfig or awsConfig must be supplied. Changes apply immediately. Do NOT use this to update an existing connection (use update_cloudflow_connection) or to trigger a flow (use trigger_cloud_flow).
update_cloudflow_connection ifMatch, connectionId, name, description, enabled, gcpConfig, awsConfig, collaborators
Updates a CloudFlow connection immediately using connectionId and the last observed ETag (ifMatch). A stale ETag fails with 412. Other fields are optional; supplied gcpConfig/awsConfig and collaborators replace their stored values wholesale. Retain exactly one owner when replacing collaborators. At most one provider config may be supplied. Serialize updates because the API version check is not atomic with the write. If an HTTP failure has only generic text, do not infer its status or retry automatically.
list_cloudflow_templates maxResults, pageToken
Use this when the user wants to see the catalogue of available CloudFlow templates (read-only blueprints they can build a flow from). Returns a cursor-paginated list of templates with their id, name, description, and instructions. Do NOT use this to view a single template's details (use get_cloudflow_template) or to trigger a flow (use trigger_cloud_flow).
get_cloudflow_template templateId
Use this when the user wants to view the details of a specific CloudFlow template by its ID, including its name, description, and configuration instructions. Do NOT use this to list all templates (use list_cloudflow_templates) or to trigger a flow (use trigger_cloud_flow).
refine_cloudflow question, flowId, conversationId
Refines an existing CloudFlow using natural language. Streams progress and returns the answer and conversationId; normally no flowId is returned. A plan or clarification question may save nothing. Reusing conversationId continues that conversation. Generated codeNode code can pass validation and fail silently or error at run time, so a successful build shows only that a draft was saved. export_cloudflow_flow returns the saved code; a completed test run's per-node output shows whether it works. Test runs execute real actions, including on drafts. Full CloudFlow authoring guide: MCP resource doit://docs/cloudflow-authoring.
build_cloud_flow question, conversationId
Use this when the user wants to build a brand-new CloudFlow automation from scratch using natural language. Creates the draft before planning, so flowId can be returned even if the builder stops early. Streams progress and returns flowId, conversationId, the answer, and any build steps. Reusing conversationId continues the conversation but this endpoint still creates a new draft; refine_cloudflow targets an existing flow. Generated codeNode code can pass validation and fail silently or error at run time, so a successful build shows only that a draft was saved. export_cloudflow_flow returns the saved code; a completed test run's per-node output shows whether it works. Test runs execute real actions, including on drafts. Full CloudFlow authoring guide: MCP resource doit://docs/cloudflow-authoring.
list_organizations pageToken
Use this when the user wants to see the organizations in their DoiT account. Returns a list of organizations. Do NOT use this for listing users (use list_users) or platforms (use list_platforms).
list_platforms pageToken
Use this when the user wants to see the support-ticket platform catalog. Returns platform IDs and display names for create_ticket; this catalog does not list connected cloud accounts. Do NOT use this for cloud incidents (use get_cloud_incidents) or products (use list_products).
list_users pageToken
Use this when the user wants to see users in their DoiT organization or check who has access. Includes active users and pending invitations. Returns roleId values, which resolve to role details in list_roles. Do NOT use this for listing roles (use list_roles) or validating the current user (use validate_user).
update_user id, firstName, lastName, jobFunction, phone, phoneExtension, language, roleId
Use this when the user wants to update a user's information such as name, job function, phone, language, or role. Changes apply immediately. Do NOT use this for inviting new users (use invite_user) or listing users (use list_users).
invite_user email, roleId, organizationId
Use this when the user wants to invite a new person to the organization. The invitation email is sent immediately. Omitted roleId defaults to Support User. The email domain must be permitted by the customer's invitation policy; existing users and pending invitees cannot be re-invited. Do NOT use this for updating existing users (use update_user) or listing users (use list_users).
list_roles pageToken
Use this when the user wants to see available roles in their DoiT organization. Returns a list of roles with permissions. Do NOT use this for listing users (use list_users) or organizations (use list_organizations).
list_products platform, pageToken
Use this when the user wants to see the support-ticket product catalog for one platform. Returns product IDs, display names, and platform IDs. create_ticket uses the product displayName and the platform ID. Requires platform: the full catalog spans several hundred products across all platforms and exceeds the MCP result size limit, so products are listed one platform at a time. Customers do not see private products. This catalog does not list products purchased by the account. Do NOT use this for cloud incidents (use get_cloud_incidents) or platforms (use list_platforms).
list_labels maxResults, pageToken, filter, sortBy, sortOrder
Use this when the user wants to see their DoiT console labels for organizing reports, budgets, alerts, allocations, metrics and annotations. Returns label metadata; these are not cloud resource labels. Do NOT use this for annotations (use list_annotations) or label assignments (use get_label_assignments).
get_label id, name
Use this when the user wants to view details of a specific DoiT console label (not a cloud resource label). Accepts either the label ID or a case-insensitive partial name. Name lookup searches only the first 200 labels; multiple matches return an error listing names, and id takes precedence. Do NOT use this for listing all labels (use list_labels) or annotations (use list_annotations).
create_label name, color
Use this when the user wants to create a new DoiT console label for organizing analytics objects. Names must be unique within the customer. Requires Cloud Analytics Admin. Changes apply immediately. Do NOT use this for viewing existing labels (use list_labels) or annotations (use create_annotation).
update_label id, name, color
Use this when the user wants to modify an existing custom DoiT console label. Requires Cloud Analytics Admin; preset labels cannot be edited. Supports partial updates; null leaves a field unchanged. Changes apply immediately. Do NOT use this for creating new labels (use create_label) or annotations (use update_annotation).
get_label_assignments id, pageToken
Use this when the user wants to see which DoiT console objects are assigned to a label. Returns objectId and objectType pairs, without object details. Do NOT use this for viewing label details (use get_label) or allocations (use list_allocations).
assign_objects_to_label id, add, remove
Use this when the user wants to assign or unassign DoiT console objects (reports, budgets, alerts, allocations, metrics or annotations) to a custom console label. Preset labels cannot be assigned. Requires edit permission on every object; validation is all-or-nothing. add/remove modify associations, not a replacement list. Changes apply immediately. Do NOT use this for creating labels (use create_label) or viewing assignments (use get_label_assignments).
list_folders maxResults, pageToken
Use this when the user wants to see their Cloud Analytics folders, which organize reports and allocations into a hierarchy. Returns a list of folders with their metadata. Do NOT use this for listing reports (use list_reports) or labels (use list_labels).
get_folder id, name
Use this when the user wants to view details of a specific Cloud Analytics folder. Accepts either the folder ID or a case-insensitive partial name. Name lookup searches only the first 200 folders; multiple matches return an error listing names, and id takes precedence. Do NOT use this for listing all folders (use list_folders) or viewing reports (use get_report_config).
create_folder name, description, parentFolderId
Use this when the user wants to create a new Cloud Analytics folder to organize reports and allocations. A duplicate sibling name or invalid parent is rejected. Changes apply immediately. Do NOT use this for creating reports (use create_report) or labels (use create_label).
update_folder id, name, description, parentFolderId
Use this when the user wants to rename, re-describe, or move (reparent) an existing Cloud Analytics folder. Changes apply immediately. Auto-renaming on collision applies only when moving to another parent; a rename that collides with a sibling is rejected. Invalid parents and moves into self or a descendant are rejected. Do NOT use this for creating new folders (use create_folder) or updating reports (use update_report).
list_themes pageToken
Use this when the user wants to see the custom color themes defined for their account, which control the colors applied to Cloud Analytics reports. Returns a list of themes with their metadata. Do NOT use this for listing reports (use list_reports) or labels (use list_labels).
get_theme id, name
Use this when the user wants to view details of a specific custom color theme. Accepts either the custom theme ID or a partial name (case-insensitive) across all custom themes. Multiple matches return an error listing names; id takes precedence. Preset themes cannot be fetched with this tool. Do NOT use this for listing all themes (use list_themes).
get_active_theme
Use this when the user wants to know which color theme is currently active for the authenticated user (the theme applied to Cloud Analytics reports). Returns the active theme id; the reserved sentinel "default" means no custom or preset theme is selected and the built-in default is in use. Do NOT use this to list all themes (use list_themes) or to fetch a specific theme by id (use get_theme).
set_active_theme themeId
Use this when the user wants to change the authenticated user’s active Cloud Analytics color theme. Accepts a custom theme ID or supported preset ID or the sentinel "default" to revert to the built-in default. Changes apply immediately. Requires Cloud Analytics edit permission. Create new themes with create_custom_theme. Do NOT use this to retrieve the current active theme (use get_active_theme) or to update theme colors (use update_theme).
update_theme id, name, newName, primaryColor, colors
Use this when the user wants to modify an existing custom color theme — rename it, change its primary color, or update its color palette. Accepts either the theme ID or a partial name match across all custom themes; ambiguous matches return an error, and id takes precedence. Requires Cloud Analytics Admin; preset themes cannot be updated. Changes apply immediately to all users who have this theme active. Omitted fields are preserved; colors replaces both palettes. Create new themes with create_custom_theme; change the current user's active theme with set_active_theme.
get_aws_account accountID
Use this when the user wants the CloudConnect details of a specific connected AWS account, such as its IAM role ARN, S3 bucket for real-time data, and which DoiT features are enabled or supported. Requires the 12-digit AWS account ID. Do NOT use this for Google Cloud or Azure accounts.
get_cloud_connect_supported_features accountID
Use this when the user wants to know which DoiT CloudConnect features a connected AWS account supports and whether the account currently has the required permissions for each feature. Accepts a 12-digit AWS account ID, including AWS standalone accounts. Azure tenant IDs are unsupported. Returns the list of supported features with their permission status.
list_datahub_datasets pageToken
Use this when the user wants to see available DataHub datasets. Returns a list of datasets with metadata. Do NOT use this for billing data (use run_query) or assets (use list_assets).
get_datahub_dataset name
Use this when the user wants to view details of a specific DataHub dataset by its name. Returns dataset metadata, including the schemaTemplate identifier when present, rather than the full column schema. Do NOT use this for listing all datasets (use list_datahub_datasets) or cost queries (use run_query).
create_datahub_dataset name, description
Use this when the user wants to create a new DataHub dataset. Changes apply immediately. Do NOT use this for viewing datasets (use list_datahub_datasets) or sending events (use send_datahub_events).
update_datahub_dataset name, description, displayName, logoName
Use this when the user wants to modify an existing DataHub dataset's description, displayName, or logoName. The dataset name identifies the dataset and cannot be changed. Omitted fields are kept; empty strings clear the supplied fields. At least one updatable field is required. Changes apply immediately. Do NOT use this for creating datasets (use create_datahub_dataset) or listing datasets (use list_datahub_datasets).
send_datahub_events events
Use this when the user wants to send DataHub events for ingestion (1–50,000 events per call). Each event requires a provider name and an RFC 3339 timestamp, and can optionally include dimensions and metrics. The API validates the entire request atomically, including duplicate event IDs anywhere in the request; a validation failure rejects the whole request. Success means accepted for asynchronous processing, not that records are already queryable. No processing latency is guaranteed. Do NOT use this for creating datasets (use create_datahub_dataset) or viewing datasets (use list_datahub_datasets).
find_cloud_diagrams resources
Use this when the user wants to find architecture diagrams or cloud infrastructure diagrams. Matches cloud resource IDs (cld_id or props.id) and returns diagram viewer URLs and image URLs. Creates a sheet filter and queues image rendering. Do NOT use this for cost analysis (use run_query) or incidents (use get_cloud_incidents).
get_cloud_diagrams_stats start, end, pageToken
Use this when the user wants activity statistics for their cloud infrastructure diagrams over a time period — node create/update/delete change counts grouped by cloud service, plus each diagram's import/sync state. Useful for change auditing and drift detection. Requires a start and end RFC3339 date-time.
search_cloud_diagrams query, ss_id, from, size
Use this when the user wants to search their cloud infrastructure diagrams and components by name or property. Returns matching diagram layers (scheme), components, and components matched by property value (prop). ss_id scopes only component and prop results, not scheme results. from/size page each category independently. Do NOT use this for cost analysis (use run_query) or incidents (use get_cloud_incidents).
get_cloud_diagram_cost_snapshot layerId, startDate, endDate, interval
Use this when the user wants a cost snapshot for a specific cloud infrastructure diagram layer over a time period — the API-reported total, trendingPct as a percentage (25 means 25%, null when no prior value is available), the top five resources and top five services by cost, and the last twelve trend buckets. Both dates are inclusive. Requires the diagram layer ID and a startDate/endDate (YYYY-MM-DD). Do NOT use this for account-wide cost analysis (use run_query) or budgets (use list_budgets).
get_cloud_diagram_resource_relationships layerId, resourceId, direction, depth, kind
Use this when the user wants to understand how a specific resource in a cloud infrastructure diagram is connected to other resources — its upstream/downstream edges and optional group membership (only when kind is group_members or both). Returns the anchor resource plus up to 200 relations with their type and hop distance; truncated is true when more than 200 relations exist. Requires the diagram layer ID and the resource ID. Do NOT use this for cost analysis (use get_cloud_diagram_cost_snapshot or run_query).
list_cloud_diagram_activity_groups ss_id, limit, offset, tags, pageToken
Use this when the user wants the activity history of a cloud diagram layer. Without tags, returns ALARM, COMMIT, EVENT, and SNAPSHOT activity groups for the given layer (ss_id), ordered by timestamp descending; snapshot groups reference a snapshot and contain their individual activity records. Supplying non-empty tags restricts results to snapshots. Page with offset/limit and filter with tags. Do NOT use this for cost analysis (use run_query) or incidents (use get_cloud_incidents).
list_cloud_diagram_node_activities ss_id, nodeId, limit, offset, pageToken
Use this when the user wants the change history of a single component node in a cloud diagram layer. Returns individual activity records (NODE_CREATE/NODE_UPDATE/NODE_DELETE) for the given node (ss_id + nodeId), ordered by timestamp descending, each including user as the ID of the user who made the change. Page with offset/limit. Do NOT use this for cost analysis (use run_query) or incidents (use get_cloud_incidents).
get_cloud_diagram_components scheme_ids, layer_ids, include_components, skip_empty, pageToken
Use this when the user wants to discover all cloud infrastructure diagrams and their layers (statussheets), or to look up layer IDs needed for other diagram endpoints. With no filters, returns accessible application and infrastructure diagrams with layer metadata and no component data. Returns maps keyed by diagram and layer IDs. Selectors must belong to diagrams accessible to the authenticated customer and user. scheme_ids and the diagrams owning layer_ids are combined, not intersected. The layer IDs required by other diagram tools come from this tool. Optionally filter by diagram IDs (scheme_ids) or layer IDs (layer_ids), and set include_components=true to load projected component maps for up to five requested or resolved layers. Larger selections return an error before loading components; discover metadata first, then request batches of at most five layer_ids. Do NOT use this for cost analysis (use run_query) or diagram search (use search_cloud_diagrams).
list_budgets maxResults, pageToken, filter, name, minCreationTime, maxCreationTime
Use this when the user wants to see their cloud spending budgets or check budget status. Returns a paginated list of budgets with names, amounts, and utilization. Do NOT use this for cost analysis (use run_query) or spending alerts (use list_alerts).
get_budget id, name
Use this when the user wants to view the details and current utilization of a specific budget. Accepts either the budget ID or a partial name (case-insensitive). Do NOT use this for listing all budgets (use list_budgets) or cost analysis (use run_query).
create_budget name, amount, currency, type, timeInterval, startPeriod, endPeriod, description, usePrevSpend, growthPerPeriod, metric, public, scopes, scope, collaborators, alerts, recipients, recipientsSlackChannels, seasonalAmounts
Use this when the user wants to create a new cloud budget with spending limits and alert thresholds. Requires name, currency, type, startPeriod, and exactly one of scope/scopes. Requires amount unless usePrevSpend is true, endPeriod for fixed budgets, and timeInterval for recurring budgets (ignored for fixed). Changes apply immediately. Do NOT use this for viewing existing budgets (use list_budgets or get_budget) or creating alerts (use create_alert).
update_budget name, amount, currency, type, timeInterval, startPeriod, endPeriod, description, usePrevSpend, growthPerPeriod, metric, public, scopes, scope, collaborators, alerts, recipients, recipientsSlackChannels, seasonalAmounts, id
Use this when the user wants to modify an existing budget. Supports partial updates; omitted fields retain their current values. Setting type to fixed requires endPeriod; setting type to recurring requires timeInterval and forbids endPeriod; setting usePrevSpend to false requires amount. Changes apply immediately. Do NOT use this for viewing budgets (use list_budgets) or creating new budgets (use create_budget).
list_annotations maxResults, pageToken, filter, sortBy, sortOrder
Use this when the user wants to see calendar annotations or notes on cost data. Returns a list of annotations. Do NOT use this for labels (use list_labels) or alerts (use list_alerts).
get_annotation id, content
Use this when the user wants to view details of a specific annotation. Accepts either the annotation ID or a partial content match (case-insensitive) within the first 200 annotations. Multiple matches return an ambiguity error listing content; id takes precedence. Do NOT use this for listing all annotations (use list_annotations) or labels (use list_labels).
create_annotation content, timestamp, reports, labels
Use this when the user wants to add a new annotation to mark a specific date or event in cost data. Changes apply immediately. Do NOT use this for creating labels (use create_label) or alerts (use create_alert).
update_annotation id, content, timestamp, reports, labels
Use this when the user wants to modify an existing annotation. Nonempty content is required on every update. Omitted or null timestamp/reports/labels remain unchanged; reports and labels replace their lists, and [] clears them. Changes apply immediately. Do NOT use this for creating new annotations (use create_annotation) or labels (use update_label).
list_commitments maxResults, pageToken, filter, sortBy, sortOrder
Returns a paginated list of spend commitment contracts from the DoiT Commitment Manager for Google Cloud, AWS, and Azure. These are negotiated spend commitments, rather than resource usage commitments; AWS agreements may be called Enterprise Discount Programs (EDP).
get_commitment id
Returns details of a specific spend commitment contract for Google Cloud, AWS, or Azure, identified by its ID. Includes the full breakdown of commitment periods, per-period contracted values, and current spend attainment against the committed amount.
list_account_team pageToken
Use this when the user wants to know who their DoiT account team / account managers are. Returns the list of account managers assigned to the customer, including name, email, role, and Calendly scheduling link. Do NOT use this for listing platform users (use list_users) or organizations (use list_organizations).
get_resource_permissions resourceType, resourceId
Use this when the user wants to see who a Cloud Analytics resource is shared with and at what access level. Returns the sharing settings (per-user roles and public visibility) for a specific alert, budget, report, or allocation. Requires resourceType (alerts, budgets, reports, or allocations) and resourceId. Do NOT use this to list the resources themselves (use list_alerts, list_budgets, list_reports, or list_allocations).
update_resource_permissions resourceType, resourceId, permissions, public
Use this when the user wants to change who a Cloud Analytics resource is shared with or update access levels. Replaces sharing settings with PUT for a specific alert, budget, custom report, or allocation. Requires resourceType, resourceId and a nonempty permissions list with exactly one owner. Read current permissions first and retain all intended entries. Only the current owner or a user with Ownership assignment permission may transfer ownership. Omitted public makes reports, alerts and budgets private; allocations reject omitted/null public. Do NOT use this to view current permissions (use get_resource_permissions).
ask_ava_sync question, conversationId, ephemeral
Ask DoiT AVA, DoiT's AI assistant for cloud cost and infrastructure, a question about the user's DoiT account, cloud spending, anomalies, or optimization opportunities. AVA has access to the customer's billing data, usage patterns, and DoiT-specific features. It answers DoiT and cloud-specific questions, not general-purpose ones. Complex questions may take a long time. Timeouts and application error envelopes (including HTTP 200) return MCP errors.
confirm_action approvalToken
Runs a generated DELETE operation that another tool staged and returned as `status: "approval_required"` with a summary and a one-time approval token. Intended for use after the user approves that summary. It runs exactly the saved request, including its customer scope, and returns that operation's own result or error. A token that is never confirmed expires after 5 minutes, and the staged action does not run. The token is bound to the authenticated caller and must match exactly; a different caller cannot run the action. A token runs at most once, even if execution fails. Creates and updates are not staged by the current tool registry.
create_signup_request Idempotency-Key, email, companyName, source, turnstileToken, termsConsent, customerContext
Start a Cloud Intelligence trial for an organization that is not yet a DoiT customer. These operations require no API key: they are rate limited, idempotent, and provision nothing until the organization's admin verifies the request. Agents create a request and hand the returned consentUrl to the admin; the admin verifies by entering the emailed code (or signing the request on the console), after which the tenant is provisioned and the request reports its customer id. Starts a Cloud Intelligence trial signup for the organization owning the email domain. Idempotent per Idempotency-Key: a retried create returns the same request. Nothing is provisioned until the admin verifies the request. Human channels (console, web) additionally require a Cloudflare Turnstile token and an accepted Terms of Service version; the emailed 6-digit verification code is then the consent act. Agents should present consentUrl to the organization's admin and poll getSignupRequest. Existing accounts, customer domains and previous trials receive the same pending response as new organizations. Eligibility is evaluated after mailbox verification or authenticated human consent.
get_signup_request signupRequestId, Signup-Result-Token, customerContext
Start a Cloud Intelligence trial for an organization that is not yet a DoiT customer. These operations require no API key: they are rate limited, idempotent, and provision nothing until the organization's admin verifies the request. Agents create a request and hand the returned consentUrl to the admin; the admin verifies by entering the emailed code (or signing the request on the console), after which the tenant is provisioned and the request reports its customer id. Returns the current state of a signup request. Poll at the returned pollAfterSeconds cadence; once status is completed the response carries customerId, consoleUrl and signInUrl. The response never includes the requester's email or the verification code. A browser must send the Signup-Result-Token returned by verifySignupRequest to read results after email verification. Without it, the request remains pending until its consent window expires. Agents may read results after the matching authenticated human explicitly approves their request.
verify_signup_request signupRequestId, code, customerContext
Start a Cloud Intelligence trial for an organization that is not yet a DoiT customer. These operations require no API key: they are rate limited, idempotent, and provision nothing until the organization's admin verifies the request. Agents create a request and hand the returned consentUrl to the admin; the admin verifies by entering the emailed code (or signing the request on the console), after which the tenant is provisioned and the request reports its customer id. Confirms ownership of the admin email by entering the 6-digit code that was emailed when the request was created (or resent). Success records consent and starts provisioning; the request moves to consented and then completed. Five wrong codes lock the current code; request a new one with resendSignupVerification.
create_signup_session signupRequestId, activationToken, customerContext
Start a Cloud Intelligence trial for an organization that is not yet a DoiT customer. These operations require no API key: they are rate limited, idempotent, and provision nothing until the organization's admin verifies the request. Agents create a request and hand the returned consentUrl to the admin; the admin verifies by entering the emailed code (or signing the request on the console), after which the tenant is provisioned and the request reports its customer id. Exchanges the one-time activationToken returned by verifySignupRequest for a Firebase custom token once the request is completed, so the browser that entered the code lands in the new tenant without a second email. The token is bound to the request, single-use and valid for 10 minutes. Sign in with signInWithCustomToken against tenantId, then open consoleUrl.
resend_signup_verification signupRequestId, customerContext
Start a Cloud Intelligence trial for an organization that is not yet a DoiT customer. These operations require no API key: they are rate limited, idempotent, and provision nothing until the organization's admin verifies the request. Agents create a request and hand the returned consentUrl to the admin; the admin verifies by entering the emailed code (or signing the request on the console), after which the tenant is provisioned and the request reports its customer id. Emails a fresh 6-digit code to the requester. At most three codes per request, at least 60 seconds apart.
list_alert_slack_channels nameContains, maxResults, pageToken, customerContext
Notifications triggered when cloud costs exceed defined thresholds or meet specific conditions. Lists Slack destinations eligible for Alert notifications for the authenticated customer and caller. Returns canonical identifiers for recipientsSlackChannels and integration status. User credentials include visible private channels; service accounts return public/shared channels only. Results are sorted by stable destination identity. Keep nameContains unchanged when following pageToken. This endpoint is paginated: a response with a non-empty `pageToken` has more results, returned when that value is passed as the `pageToken` parameter. A missing, null, or empty `pageToken` marks the last page.
delete_alert id, customerContext
Notifications triggered when cloud costs exceed defined thresholds or meet specific conditions. Deletes the alert specified by the Id. Two-step: a call with valid arguments deletes nothing. It returns `status: "approval_required"` with a summary of the exact request and a one-time `approvalToken`. The saved request runs only when `confirm_action` is called with that token after the user approves the summary. The token is bound to the caller and expires after 5 minutes.
delete_allocation id, customerContext
Define how costs are distributed across your organization. Deletes the allocation specified by the Id. Two-step: a call with valid arguments deletes nothing. It returns `status: "approval_required"` with a summary of the exact request and a one-time `approvalToken`. The saved request runs only when `confirm_action` is called with that token after the user approves the summary. The token is bound to the caller and expires after 5 minutes.
delete_annotation id, customerContext
Custom notes added to cost data to provide contextual information. Deletes the annotation specified by the Id. Two-step: a call with valid arguments deletes nothing. It returns `status: "approval_required"` with a summary of the exact request and a one-time `approvalToken`. The saved request runs only when `confirm_action` is called with that token after the user approves the summary. The token is bound to the caller and expires after 5 minutes.
delete_label id, customerContext
Create and manage labels to organize and categorize your cloud resources. Deletes the label specified by the Id. Two-step: a call with valid arguments deletes nothing. It returns `status: "approval_required"` with a summary of the exact request and a one-time `approvalToken`. The saved request runs only when `confirm_action` is called with that token after the user approves the summary. The token is bound to the caller and expires after 5 minutes.
create_custom_theme name, primaryColor, colors, customerContext
Configure presentation settings for Cloud Analytics reports, including custom color themes. Creates a new custom color theme. Requires Cloud Analytics Admin permission.
delete_custom_theme id, customerContext
Configure presentation settings for Cloud Analytics reports, including custom color themes. Deletes the custom theme specified by the Id. Requires Cloud Analytics Admin permission. Two-step: a call with valid arguments deletes nothing. It returns `status: "approval_required"` with a summary of the exact request and a one-time `approvalToken`. The saved request runs only when `confirm_action` is called with that token after the user approves the summary. The token is bound to the caller and expires after 5 minutes.
delete_budget id, customerContext
Track actual cloud spend against planned spend. Deletes the specified budget. Two-step: a call with valid arguments deletes nothing. It returns `status: "approval_required"` with a summary of the exact request and a one-time `approvalToken`. The saved request runs only when `confirm_action` is called with that token after the user approves the summary. The token is bound to the caller and expires after 5 minutes.
list_budget_suggestions customerContext, pageToken
AI-generated budget recommendations you can accept (link to a budget you created) or dismiss. Returns the pending AI-generated budget suggestions for your account. The set is small (a handful
of pending suggestions) and is returned in full. Each suggestion can be accepted (after you create a
matching budget via `POST /analytics/v1/budgets`) or dismissed.
accept_budget_suggestion id, budgetId, editedBeforeAccept, customerContext
AI-generated budget recommendations you can accept (link to a budget you created) or dismiss. Marks the suggestion as accepted and links it to an existing budget. Create the budget first via
`POST /analytics/v1/budgets`, then pass its `id` as `budgetId`. The budget must belong to your account.
dismiss_budget_suggestion id, reason, customerContext
AI-generated budget recommendations you can accept (link to a budget you created) or dismiss. Marks the suggestion as dismissed so it no longer appears in the pending list.
delete_folder id, customerContext
Organize Cloud Analytics resources (reports, allocations) into folders. Deletes the specified folder. All nested folders will be deleted. Any reports or allocations contained in the folder are moved to the root. Two-step: a call with valid arguments deletes nothing. It returns `status: "approval_required"` with a summary of the exact request and a one-time `approvalToken`. The saved request runs only when `confirm_action` is called with that token after the user approves the summary. The token is bound to the caller and expires after 5 minutes.
delete_report id, customerContext
Manage Cloud Analytics reports and get reports data in JSON format. Deletes the specified Cloud Analytics report. Two-step: a call with valid arguments deletes nothing. It returns `status: "approval_required"` with a summary of the exact request and a one-time `approvalToken`. The saved request runs only when `confirm_action` is called with that token after the user approves the summary. The token is bound to the caller and expires after 5 minutes.
async_run_inline dryRun, Idempotency-Key, config, customerContext
Manage Cloud Analytics reports and get reports data in JSON format. Submits an async report execution job using an inline configuration. Returns 202 immediately with a Location header pointing to the operation status endpoint. Requires the Idempotency-Key header to ensure at-most-once submission. Duplicate requests with the same config for the same customer return the existing in-flight operation. Use ?dryRun=true to validate the config without creating an operation.
get_async_operation operationId, customerContext
Manage Cloud Analytics reports and get reports data in JSON format. Returns the current status of an async report operation. Non-terminal operations (pending, running) include a Retry-After header suggesting when to poll again. This endpoint does not return result data — once status is "succeeded", fetch the result from the operation's results endpoint. When status is "failed", the error field contains an RFC 9457-shaped error. Returns 404 if the operationId does not exist, belongs to a different tenant, or has expired.
async_run_report_by_id id, dryRun, timeRange, startDate, endDate, Idempotency-Key, customerContext
Manage Cloud Analytics reports and get reports data in JSON format. Submits an async execution job for a saved report identified by ID. Returns 202 immediately with a Location header pointing to the operation status endpoint. Requires the Idempotency-Key header. Use ?dryRun=true to validate without creating an operation.
cancel_async_operation operationId, dryRun, Idempotency-Key, customerContext
Manage Cloud Analytics reports and get reports data in JSON format. Cancels a pending or running async report operation. Already-terminal operations (succeeded, failed, canceled) are returned as-is without any state change (idempotent). Returns 404 if the operationId does not exist or belongs to a different tenant. Use ?dryRun=true to validate the request without modifying any state.
get_async_operation_results operationId, fileOutput, customerContext
Manage Cloud Analytics reports and get reports data in JSON format. Returns the result of a succeeded async report operation, including report metadata (id, reportName, owner, type, createTime, updateTime, urlUI) when the operation was started against a saved report — the same shape as the sync GetReportResponse, instead of requiring a second call to GET /analytics/v1/reports/{id}/config for it. Returns 404 if the operationId does not exist, has expired, or belongs to a different tenant. Returns 425 Too Early if the operation has not yet reached a terminal state — poll the operation status endpoint, which returns its own Retry-After guidance, until it succeeds. Returns 422 if the operation terminated as failed or canceled. Returns 413 if the result is too large to return inline — request it again with `fileOutput=json`, or narrow the report. The poll status endpoint response does not include result data inline — this is the only endpoint that returns it.
list_widgets X-Tenant-Id, customerContext, pageToken
Beta. Read precomputed current-month cloud spend and forecast metrics. Widget results are
refreshed in the background; reading a widget returns the latest cached result and never
starts a synchronous refresh. Use the list operation to discover the complete catalogue. Returns the complete catalogue of preset widgets supported by the Analytics Widgets API.
Each entry contains a customer-specific `id` and a stable customer-facing `alias`; either
value can be passed to the get-widget operation. An `id` can vary by customer or service
tier, while an `alias` remains stable.

Authenticate with `Authorization: Bearer <token>`. Omit `X-Tenant-Id` when the credential
resolves to one customer. When the credential can act on multiple customers, send
`X-Tenant-Id: <tenant-id>` to select the customer. A `403` response means the caller or
selected customer is not authorized or entitled to use this API.
get_widget widgetId, X-Tenant-Id, customerContext
Beta. Read precomputed current-month cloud spend and forecast metrics. Widget results are
refreshed in the background; reading a widget returns the latest cached result and never
starts a synchronous refresh. Use the list operation to discover the complete catalogue. Returns the latest precomputed cached result for one widget. This read does not refresh or
recompute the widget synchronously.

Use either the customer-specific `id` returned by the list-widgets operation or one of these
stable aliases:

- `current-month-cloud-spend` — total cloud spend accumulated during the current calendar month.
- `current-month-cloud-forecast` — forecast of total cloud spend for the current calendar month.

The two concrete alias requests are:

- `GET https://api.doit.com/analytics/v1/widgets/current-month-cloud-spend`
- `GET https://api.doit.com/analytics/v1/widgets/current-month-cloud-forecast`

Authenticate with `Authorization: Bearer <token>`. Omit `X-Tenant-Id` when the credential
resolves to one customer. When the credential can act on multiple customers, send
`X-Tenant-Id: <tenant-id>` to select the customer. A `403` response means the caller or
selected customer is not authorized or entitled to use this API.

The result's `period.endTime` is exclusive. Monetary `amount` values are decimal strings and
`currency` values are ISO 4217 currency codes. `monthToMonthGrowthPercentage` and
`monthToMonthGrowthAmount` are both `null` when no comparable prior-period value exists.
list_contracts customerID, customerContext, pageToken
List and manage tenant-scoped contracts as a T1/T2 PartnerOps caller. Lists the contracts held by the specified customer. Callable by a T1/T2 PartnerOps principal for its own tenant or any descendant tenant. Read access requires the contractsViewer or contractsManager permission, plus the billing:contracts or channelops entitlement. User API tokens must include one of the permissions in their scope.
get_contract customerID, contractID, customerContext
List and manage tenant-scoped contracts as a T1/T2 PartnerOps caller. Returns the specified contract. Requires the contractsViewer or contractsManager permission. User API tokens must include one of them in their scope.
patch_anomaly id, customerFeedback, customerContext
Monitor cost spikes in your cloud environment. Updates the review status of the anomaly identified by `{id}`, the same identifier `getAnomaly` accepts. Mirrors the review workflow available in the console: set `reviewStatus`, and when resolving, qualify it with `resolution` and optionally a `resolutionNote` (feedback reason and comment). The change is recorded in the anomaly's review history exactly like a console update, and the actor is derived from the API credentials - never from the request body.

Validation: `reviewStatus` is one of `NEEDS_REVIEW`, `UNDER_REVIEW`, `RESOLVED`; `resolution` (`ANOMALY_CONFIRMED` or `NOT_ANOMALY`) is required when `reviewStatus` is `RESOLVED` and forbidden otherwise; `resolutionNote.reason` must belong to the given `resolution` - `SECURITY_BREACH`, `MISCONFIGURATION`, `UNINTENTIONAL_PROVISIONING`, `EXPECTED_COST_SPIKE` or `ANOMALOUS_OTHER` for `ANOMALY_CONFIRMED`, and `FAULTY_ANOMALY_DETECTION_MODEL`, `INCORRECT_DATA`, `LOW_IMPACT` or `NON_ANOMALOUS_OTHER` for `NOT_ANOMALY`. Requires the `AnomaliesManager` permission and the `governance:costAnomalies` and `governance:costAnomalies:acknowledgement` entitlements.
get_anomaly_explanation id, customerContext
Monitor cost spikes in your cloud environment. Returns a likely-cause explanation for the specified anomaly, alongside the deterministic facts and evidence references it was generated from. The explanation itself is AI-generated; it is always returned separately from the deterministic `facts`, and is marked as such in the `explanation` object. Generation is ephemeral: no conversation is persisted on your behalf. Requires the `governance:costAnomalies` entitlement and the `AnomaliesViewer` permission.
create_asset type, mode, accountName, rootEmail, customerContext
Manage cloud resources or services in your cloud environment. Creates a new asset.
datahub_events_csv_file provider, file, schemaTemplate, periodStart, periodEnd, replace, customerContext
Ingest third-party cost, usage, and metric-based data for analysis. Sends a batch of events to DataHub using a CSV file, either uncompressed or compressed in ZIP or GZ format. It may take up to 15 minutes for the data to become available in the DoiT console.
datahub_import_provider_records provider, schemaTemplate, periodStart, periodEnd, replace, records, customerContext
Ingest third-party cost, usage, and metric-based data for analysis. Sends provider-native records as JSON instead of a CSV file. The records use the same columns as the template's CSV (currently `anthropic-spend-report@1` only) and are validated, normalized and booked on `periodEnd` exactly like a CSV upload. It may take up to 15 minutes for the data to become available in the DoiT console.
delete_datahub_events_by_filter dataset, eventIds, startTime, endTime, customerContext
Ingest third-party cost, usage, and metric-based data for analysis. Deletes specific events using filters. Note that the two filters, `eventIds` and `time ranges`, are mutually exclusive.
delete_datahub_datasets datasets, customerContext
Ingest third-party cost, usage, and metric-based data for analysis. Deletes one or more DataHub datasets and all their associated data. Two-step: a call with valid arguments deletes nothing. It returns `status: "approval_required"` with a summary of the exact request and a one-time `approvalToken`. The saved request runs only when `confirm_action` is called with that token after the user approves the summary. The token is bound to the caller and expires after 5 minutes.
delete_datahub_dataset name, customerContext
Ingest third-party cost, usage, and metric-based data for analysis. Deletes a specific DataHub dataset. Two-step: a call with valid arguments deletes nothing. It returns `status: "approval_required"` with a summary of the exact request and a one-time `approvalToken`. The saved request runs only when `confirm_action` is called with that token after the user approves the summary. The token is bound to the caller and expires after 5 minutes.
export_datahub_dataset_records name, format, startTime, endTime, maxResults, pageToken, customerContext
Ingest third-party cost, usage, and metric-based data for analysis. Returns one page of the live records of a DataHub dataset, as CSV (default) or as newline-delimited JSON in the same shape as the `/datahub/v1/events` payload.

A time window is required: `startTime` is inclusive, `endTime` is exclusive, and the window must not exceed 366 days. Rows are ordered by their event time and event id, and pages never overlap: read the `X-Next-Page-Token` response header and pass it back as `pageToken` until the header is absent. `X-Row-Count` carries the number of rows in the page. Only live rows are returned; rows deleted through the console or the delete endpoints are excluded.

Every row starts with five provenance columns (`event_id`, `batch`, `source`, `export_time`, `updated_by`) followed by the dataset's business columns in the CSV ingest vocabulary (`usage_date`, `fixed.<key>`, `label.<key>`, `project_label.<key>`, `system_label.<key>`, `metric.<type>`). The business columns are computed from the rows of each page, so consecutive pages can have different label and metric columns; union the headers when concatenating pages. Datasets created with the FOCUS schema template use the FOCUS column names instead and can only be exported as CSV.

To re-import an export into another dataset, drop the `batch`, `source`, `export_time` and `updated_by` columns, and either drop `event_id` or rename it to `id`. This tool returns a JSON envelope with data (the unchanged CSV or JSONL page body) and pageToken (from X-Next-Page-Token). Pass a non-empty pageToken back as pageToken to advance; null means the final page. MCP does not expose the HTTP headers directly.
get_invoice_pdf id, customerContext
Access your current and historical billing documents. Returns the PDF document for the invoice specified by the invoice number, including a cancelled invoice or credit memo. The response body is the PDF bytes.
The PDF is returned decompressed by default. It is only gzip-compressed when the request advertises `Accept-Encoding: gzip`, in which case the response carries `Content-Encoding: gzip` with a `Content-Length` for the compressed bytes and must be decoded as a standard `Content-Encoding: gzip` body. HTTP clients whose stack negotiates the encoding automatically also decode it automatically; a client that sets the header by hand may instead receive the raw gzip bytes (for example `curl` decodes only with `--compressed`). Omit `Accept-Encoding` (or send `identity`) to always receive the decompressed PDF. A very large invoice may exceed the decompressed-response size limit and respond with `500`; request it with `Accept-Encoding: gzip` to stream the compressed bytes without that limit.
A `404` is returned both when no invoice with the given number exists for the authenticated tenant and when the invoice exists but has no PDF document available.
get_billing_explainer_per_payer invoiceMonth, X-Tenant-Id, customerContext
Explain month-over-month changes in invoiced cloud costs. Returns the invoiced cost changes for each payer in the authenticated tenant.
get_entity_invoice_explainer billingProfileId, invoiceNumber, X-Tenant-Id, customerContext
Explain month-over-month changes in invoiced cloud costs. Returns invoiced cost changes for an invoice owned by the specified billing profile in the authenticated tenant.
list_service_quotas cloudProvider, minUtilizationPercent, maxResults, pageToken, customerContext
Monitor cloud service quota usage across connected accounts and projects. Returns the latest service quota usage snapshots collected by DoiT for the authenticated customer.
Results include only quotas retained by DoiT's monitoring collectors and are not a complete or live inventory from the cloud providers.
Results are sorted by utilization percentage in descending order. This endpoint is paginated: a response with a non-empty `pageToken` has more results, returned when that value is passed as the `pageToken` parameter. A missing, null, or empty `pageToken` marks the last page.
list_contract_templates customerContext, pageToken
Manage contract templates for PartnerOps resellers (T1/T2). Lists contract templates owned by the authenticated tenant (from the bearer token). Requires ContractTemplatesAdmin and the `channelops:contracts:templates` entitlement.
get_contract_template templateID, customerContext
Manage contract templates for PartnerOps resellers (T1/T2). Returns a single contract template owned by the authenticated tenant (from the bearer token). Requires ContractTemplatesAdmin and the `channelops:contracts:templates` entitlement.
assign_contract_template templateID, customerIds, customerContext
Manage contract templates for PartnerOps resellers (T1/T2). Applies a contract template owned by the authenticated tenant (from the bearer token) to child customers, creating a draft contract for each. Drafts have no billing effect until each one is completed and activated. The template must not be archived, and the current date must fall within its eligibility window. Requires ContractTemplatesAdmin and the `channelops:contracts:templates` entitlement.
list_customer_notifications customerContext, pageToken
Personal notification configs for a user, and organization-wide cost anomaly notifications. Returns every organization-wide cost-anomaly notification config for the tenant as an unpaginated `{ "items": [...] }` array. Items are ordered by `createTime` ascending, then by `id` ascending. Configs that are not cost-anomaly configs are omitted.
Requires the `settings` (Manage settings) permission. The API does not offer the Console's read-only Organization list; a caller without this permission receives `403`.
create_customer_notification dryRun, Idempotency-Key, name, destinations, settings, customerContext
Personal notification configs for a user, and organization-wide cost anomaly notifications. Creates an organization-wide cost-anomaly notification config. Destinations are `emails`, `slackChannels`, and `teamsChannels` arrays. The response returns those three arrays.
Requires the `settings` (Manage settings) permission.
The `Idempotency-Key` header is required. The same key and the same body replay the stored success. A different body is `422` `idempotency_key_reused`. A `5xx` response may leave the key reserved, so a retry returns `409` until the original request is released or about 30 minutes pass. `dryRun=true` does not reserve the key. With `dryRun=true` the request goes through the same checks as a real create and returns `200` with `X-Dry-Run: true` and null identity fields (`id`, `createTime`, `updateTime`, `etag`) without creating anything.
Reads are immediately consistent. After a `201`, a GET of that notification returns the created config.
get_customer_notification id, customerContext
Personal notification configs for a user, and organization-wide cost anomaly notifications. Returns one organization-wide cost-anomaly notification config. Requires the `settings` (Manage settings) permission. The API does not offer the Console's read-only Organization list; a caller without this permission receives `403`. An id that is not a cost-anomaly config for this tenant returns `404`, without saying whether it exists elsewhere.
delete_customer_notification id, dryRun, If-Match, Idempotency-Key, customerContext
Personal notification configs for a user, and organization-wide cost anomaly notifications. Permanently deletes an organization-wide cost-anomaly notification config. Requires the `settings` (Manage settings) permission and the `If-Match` header. The `Idempotency-Key` header is required. The same key, the same body, and the same `If-Match` replay the stored success. A different body or a different `If-Match` is `422` `idempotency_key_reused`. A `5xx` response may leave the key reserved, so a retry returns `409` until the original request is released or about 30 minutes pass. `dryRun=true` does not reserve the key. With `dryRun=true` the request goes through the same checks as a real delete without deleting anything. Deleting an id that is not a cost-anomaly config, or that no longer exists, returns `404`.
Reads are immediately consistent. After a completed delete (`204` without `X-Dry-Run: true`), a GET of that notification returns `404`. Two-step: a call with valid arguments deletes nothing. It returns `status: "approval_required"` with a summary of the exact request and a one-time `approvalToken`. The saved request runs only when `confirm_action` is called with that token after the user approves the summary. The token is bound to the caller and expires after 5 minutes.
update_customer_notification id, dryRun, If-Match, Idempotency-Key, name, destinations, settings, customerContext
Personal notification configs for a user, and organization-wide cost anomaly notifications. Partially updates an organization notification with an `application/merge-patch+json` body (RFC 7396). Omitted fields stay unchanged; nested objects are deep-merged; nested arrays (for example `allocations.ids`) are replaced when present. When `destinations` is present it replaces the whole destinations object; the nested `emails`, `slackChannels`, and `teamsChannels` arrays are the full lists. Optional resend fields are cleared with `null` (for example `"settings":{"costAnomalies":{"repeat":null}}`). `minSeverity`, `minExcessCost`, and `allocations` cannot be cleared.
Requires the `settings` (Manage settings) permission and the `If-Match` header. The `Idempotency-Key` header is required. The same key, the same body, and the same `If-Match` replay the stored success. A different body or a different `If-Match` is `422` `idempotency_key_reused`. A `5xx` response may leave the key reserved, so a retry returns `409` until the original request is released or about 30 minutes pass. `dryRun=true` does not reserve the key. With `dryRun=true` the request goes through the same checks as a real update and returns the notification as it would look after the patch, without changing anything.
Reads are immediately consistent. After a completed update, a GET of that notification returns the new state.
get_customer customerId, customerContext
Read and update your organization's general settings. Returns the customer, including its general settings and contact info, scoped to `{customerId}`. `{customerId}` must match the customer resolved from the bearer token; a token scoped to a different customer gets `403`, even if that customer would otherwise be reachable through a reseller/MTS relationship. Requires the `Settings` permission.
update_customer customerId, urlSlug, settings, contact, customerContext
Read and update your organization's general settings. Partially updates the general settings and contact info of the customer identified by `{customerId}`. `{customerId}` must match the customer resolved from the bearer token; a token scoped to a different customer gets `403`, even if that customer would otherwise be reachable through a reseller/MTS relationship. Requires the `Settings` permission; updating `allowedInviteDomains` additionally requires the `UsersManager` permission.

The request body must use `application/merge-patch+json` (RFC 7396): an omitted field leaves the current value unchanged, and an explicit `null` also leaves it unchanged, except for `urlSlug`, where an explicit empty string removes the customer's active URL slug rather than leaving it unchanged.

Fields are nested exactly as `getCustomer` returns them - `currency` and `allowedInviteDomains` under `settings`, `emails` under `contact` - so every value is read and written at the same path. `settings.currency` accepts only the codes listed in the schema and cannot be cleared; `allowedInviteDomains` and `contact.emails` are cleared with an empty array.
list_shared_payer_account_mappings customerId, managementAccount, customerContext, pageToken
Link and manage shared-payer account mappings for PartnerOps resellers (T2), reassigning AWS accounts under a shared payer to T3 end customers. Lists the AWS accounts under the reseller's linked shared payers, together with the end customer each is currently mapped to. `{customerId}` (the reseller itself) means the account has not yet been reassigned to an end customer. Each payer's own management account is included alongside its member accounts. Requires the `SharedPayersAdmin` permission and the AWS integration (`pdi:aws`) entitlement. User API tokens must include the permission in their scope.
list_shared_payer_management_accounts customerId, customerContext, pageToken
Link and manage shared-payer account mappings for PartnerOps resellers (T2), reassigning AWS accounts under a shared payer to T3 end customers. Lists the reseller's linked shared-payer management (payer) accounts that have completed onboarding, for use in an account picker. Requires the `SharedPayersAdmin` permission and the AWS integration (`pdi:aws`) entitlement. User API tokens must include the permission in their scope.
list_shared_payers customerId, customerContext, pageToken
Link and manage shared-payer account mappings for PartnerOps resellers (T2), reassigning AWS accounts under a shared payer to T3 end customers. Lists every shared payer linked to the reseller, including ones still mid-onboarding, with per-payer counts of linked and mapped accounts. Requires the `SharedPayersAdmin` permission and the AWS integration (`pdi:aws`) entitlement. User API tokens must include the permission in their scope.
reassign_shared_payer_account customerId, assetId, customerContext
Link and manage shared-payer account mappings for PartnerOps resellers (T2), reassigning AWS accounts under a shared payer to T3 end customers. Reassigns one AWS account under a linked shared payer - a member account or the payer's own management account - to the reseller itself or to a T3 end customer within its hierarchy. Requires the `SharedPayersAdmin` permission and the AWS integration (`pdi:aws`) entitlement. User API tokens must include the permission in their scope.
list_customer_subscriptions maxResults, pageToken, customerContext
Create, view, update, and delete dashboard, report, commitment, alert, and budget subscriptions. This is not Organization-wide notifications. Lists dashboard, report, commitment, alert, and budget subscriptions visible to the caller. Items are ordered by `id` ascending. Nested `subscriptions` are ordered by `type` ascending. `maxResults` defaults to 50. An omitted value, `0`, or a value above 500 is treated as 50. A non-integer `maxResults` is `400`. Pass the response `pageToken` to request the next page. That field is absent when there are no more pages. A `pageToken` that matches no id is `400`. `rowCount` is the total number of subscriptions visible to the caller. That value is the same on every page.
One subscription record is one item. Its dashboard, report, and commitment subscriptions are nested entries that share `recipients`. An alert is one item. A budget is one item. An alert or a budget never shares an item with a scheduled subscription. Draft budgets are omitted. An alert is included only when the tenant has the Governance Alerts entitlement and the caller can view that alert. When the entitlement is missing or cannot be confirmed, the list contains no alerts. An alert with no recipients is still included.
One id namespace is shared by a scheduled subscription, an alert, and a budget. Create accepts only `dashboard-subscription`, `report-subscription`, and `commitment-manager`. Those scheduled types are the resource: create, recipient replace, subscription-set replace, and delete. GET returns alerts and budgets so a caller can see who receives them. Create, delete, and full management of those resources, including recipients, stay on `PATCH /analytics/v1/alerts/{id}` and `PATCH /analytics/v1/budgets/{id}`. PATCH on this API may replace those same recipient lists. DELETE of an alert or a budget returns `409` `customers.entry_is_resource`.
Manage Settings users can view every subscription. A scheduled subscription is also visible to its creator when the caller has the permission for that type, to an owner, editor, or viewer of its source, and to one of its email recipients. Users Manager alone grants no access. A service-account token is evaluated from its own permissions and does not inherit the permissions of the subscription creator. A caller who can see only some nested subscriptions still receives that record; `subscriptions` omits the entries they cannot see. This endpoint is paginated: a response with a non-empty `pageToken` has more results, returned when that value is passed as the `pageToken` parameter. A missing, null, or empty `pageToken` marks the last page.
create_customer_subscription dryRun, Idempotency-Key, recipients, subscriptions, customerContext
Create, view, update, and delete dashboard, report, commitment, alert, and budget subscriptions. This is not Organization-wide notifications. Creates one dashboard, report, or commitment subscription with one, two, or three scheduled subscriptions: `dashboard-subscription`, `report-subscription`, and `commitment-manager`. Create does not accept `alert` or `budget`. Each type appears at most once. The response is that subscription. Nested entries are ordered by `type`. `name` is `{source name} (Dashboard Subscription)`, `(Report Subscription)`, or `(Commitment Subscription)`, taken from the nested type that sorts first. `createdBy` is the caller.
`subscriptions` must contain one to three entries. A repeated type is `422` `customers.duplicate_subscription_type`. An `alert` or `budget` type, an empty `subscriptions` array, or any server-set field (`id`, `name`, `createdBy`, `etag`, `permissions`, `source.name`, `schedule.nextAt`) is `422` `validation_failed`. `source.id` must belong to the tenant. A missing or foreign source is `422` `validation_failed`. Recipient and schedule value rules are on `CustomerSubscriptionRecipients` and `CustomerSubscriptionScheduleInput`.
The caller must be allowed to manage every sent type: a DoiT employee, Manage Settings, or the permission for that type. Users Manager alone grants no access.
`Idempotency-Key` is required. The same key, caller, customer, method, path, and body replays the stored success response. A different body for that key is `422` `idempotency_key_reused`. A missing key is `400` `idempotency_key_required`. While the first request is in progress, another request with that key is `409` `idempotency_request_in_progress`. A `4xx` response releases the key. `dryRun=true` does not reserve the key. It returns the would-be subscription with `id` and `etag` null and writes nothing.
Reads are immediately consistent. After a `201`, a GET of that id returns the created subscription. The first delivery can take up to an hour.
get_customer_subscription id, customerContext
Create, view, update, and delete dashboard, report, commitment, alert, and budget subscriptions. This is not Organization-wide notifications. Returns one dashboard, report, commitment, alert, or budget subscription. The body is the same object as a list item. `schedule` is omitted on an alert or a budget.
Create does not accept `alert` or `budget`. GET returns them. PATCH may replace their recipient lists, which writes the same lists as `PATCH /analytics/v1/alerts/{id}` and `PATCH /analytics/v1/budgets/{id}`. Create, delete, and full management stay on those APIs. DELETE of an alert or a budget returns `409` `customers.entry_is_resource`.
Manage Settings users can view the subscription. A scheduled subscription is also visible to its creator when the caller has the permission for that type, to an owner, editor, or viewer of its source, and to one of its email recipients. Users Manager alone grants no access. A service-account token is evaluated from its own permissions and does not inherit the permissions of the subscription creator. Keys the caller cannot see are omitted from `subscriptions`. A missing, foreign-tenant, or non-visible subscription returns `404` without revealing whether it exists.
delete_customer_subscription id, dryRun, If-Match, Idempotency-Key, customerContext
Create, view, update, and delete dashboard, report, commitment, alert, and budget subscriptions. This is not Organization-wide notifications. Deletes a dashboard, report, or commitment subscription and every scheduled entry on it. Returns `204`. Deleting an id that is already gone also returns `204`, so a retry does not fail. `If-Match` is required when the document still exists; a stale value is `412`. An id that matches no scheduled subscription, alert, or budget returns `204` when `If-Match` is present.
Create does not accept `alert` or `budget`. GET returns them. PATCH may replace their recipient lists. DELETE of an alert or a budget returns `409` `customers.entry_is_resource`.
The caller must have `canDelete` for the subscription. Manage Settings users, a creator with the permission for each scheduled type, and a source owner or editor may delete it. Source viewers and recipients cannot. Users Manager alone grants no access. A service-account token is evaluated from its own permissions and does not inherit the permissions of the subscription creator.
`Idempotency-Key` is required. The same key, caller, customer, method, path, and `If-Match` replays the stored success response. A different `If-Match` is `422` `idempotency_key_reused`. A missing key is `400` `idempotency_key_required`. While the first request is in progress, another request with that key is `409` `idempotency_request_in_progress`. A `4xx` response releases the key. `dryRun=true` does not reserve the key and does not delete.
Reads are immediately consistent. After a completed delete, a GET of that id returns `404`. Two-step: a call with valid arguments deletes nothing. It returns `status: "approval_required"` with a summary of the exact request and a one-time `approvalToken`. The saved request runs only when `confirm_action` is called with that token after the user approves the summary. The token is bound to the caller and expires after 5 minutes.
update_customer_subscription id, dryRun, If-Match, Idempotency-Key, recipients, subscriptions, customerContext
Create, view, update, and delete dashboard, report, commitment, alert, and budget subscriptions. This is not Organization-wide notifications. Partially updates one dashboard, report, commitment, alert, or budget subscription with an `application/merge-patch+json` body (RFC 7396). The writable fields are `recipients.emails`, `recipients.slackChannels`, and, on a scheduled subscription, `subscriptions`. Any other field, `recipients: null`, or `null` for either recipient list is `422` `validation_failed`. A patch that sets neither `recipients` nor `subscriptions` is `422` `validation_failed`. A wrong `Content-Type` is `415`. Recipient and schedule value rules are on `CustomerSubscriptionRecipients`, `CustomerSubscriptionRecipientPatch`, and `CustomerSubscriptionScheduleInput`.
On a scheduled subscription, `subscriptions` is the desired set of dashboard, report, and commitment subscriptions. A type in the body and not on the document is added. A type on the document and missing from the body is removed. A type in both is updated when `source.id` or the schedule differs, and `nextAt` is recalculated. A stored entry this API does not list stays on the subscription. An empty `subscriptions` array is `422` `validation_failed`; deleting the item is `DELETE`. A repeated type is `422` `customers.duplicate_subscription_type`. An omitted `subscriptions` array leaves the nested types unchanged.
Create does not accept `alert` or `budget`. GET returns them. On an alert or a budget, `subscriptions` is accepted only when it is the one stored entry with the same `type` and `source.id` and no `schedule`. Any other array is `422` `validation_failed`. Those items never gain a dashboard, report, or commitment subscription. Alerts and budgets keep their own APIs for create, delete, and full management. This `PATCH` is the one write this API makes on them, and it may replace their recipient lists. DELETE of an alert or a budget returns `409` `customers.entry_is_resource`.
A patch whose normalized values match the stored values returns `200` and does not write. The `etag` is unchanged.
Recipient changes are allowed when the caller can edit every key on the document, including a key the response did not show. Otherwise they are allowed only when the caller removes exactly their own email: no added address, no Slack change, and no subscription-set change. Any other recipient change is `403`. The caller must be allowed to manage every added or changed scheduled type and to delete every removed type.
`If-Match` must contain the current strong ETag. The token only covers fields this API can write. `Idempotency-Key` is required. The same key, caller, customer, method, path, body, and `If-Match` replays the stored success response. A different body or `If-Match` is `422` `idempotency_key_reused`. A missing key is `400` `idempotency_key_required`. While the first request is in progress, another request with that key is `409` `idempotency_request_in_progress`. A `4xx` response rel
delete_user id, customerContext
Manage users who have access to the DoiT platform. Deletes a user. Two-step: a call with valid arguments deletes nothing. It returns `status: "approval_required"` with a summary of the exact request and a one-time `approvalToken`. The saved request runs only when `confirm_action` is called with that token after the user approves the summary. The token is bound to the caller and expires after 5 minutes.
resend_invite id, dryRun, Idempotency-Key, customerContext
Manage users who have access to the DoiT platform. Resets the invite expiry to 48 hours from now, invalidates the previous invite token (so
old email links stop working), and triggers a fresh invitation email. Works on invites in
any state including `Cancelled` — resending a cancelled invite reactivates it to `Pending`.

Returns `404` if no invite exists for the given ID (never created, or already accepted and removed).

Requires `usersManager` permission.
cancel_invite id, dryRun, Idempotency-Key, customerContext
Manage users who have access to the DoiT platform. Marks the invite as `Cancelled` and invalidates the invite token so any outstanding email
links stop working. The invite document is retained (soft cancel) — the user row remains
visible in `GET /iam/v1/users` with `inviteStatus: Cancelled`. Use `DELETE /iam/v1/users/{id}`
to fully remove the record.

Returns `404` if no invite exists for the given ID, and `409` if the invite is already cancelled.

Requires `usersManager` permission.
list_user_notifications userId, customerContext, pageToken
Personal notification configs for a user, and organization-wide cost anomaly notifications. Returns every personal notification config for the user as an unpaginated `{ "items": [...] }` array. One user has a handful of configs; there is no pagination. Items are ordered by `createTime` ascending, then by `id` ascending.
Requires Users Manager (`usersManager`). The user id must belong to the resolved tenant.
create_user_notification userId, dryRun, Idempotency-Key, name, settings, customerContext
Personal notification configs for a user, and organization-wide cost anomaly notifications. Creates a personal notification config for the user. The request body has no email address; the server copies the target user's email onto the config as the sole destination. `settings` has one property per subscribed type and must include at least one. `costAnomalies` is optional.
Requires Users Manager (`usersManager`). The user id must belong to the resolved tenant. For each subscribed type other than `costAnomalies`, the target user must also hold that type's permission. The permission id for each type is documented on `UserNotificationSettings`. If they do not, the response is `403` and the problem body includes `permission` (the permission id) and `notificationType` (the settings property). A permission lookup that fails is `503`. If the permission service does not recognize an invited user, the response is `403`.
Requires the `Idempotency-Key` header so a retried create with the same key and body returns the original response and does not create a second config. With `dryRun=true` the request goes through the same checks as a real create and returns `200` with `X-Dry-Run: true` and null identity fields (`id`, `createTime`, `updateTime`, `etag`) without creating anything.
Reads are immediately consistent. After a `201`, a GET of that notification returns the created config.
get_user_notification userId, id, customerContext
Personal notification configs for a user, and organization-wide cost anomaly notifications. Returns one personal notification config. Requires Users Manager (`usersManager`). A subscribed type stays in the response when the target user later loses that type's permission; the type permission is not re-checked on read. A notification id that does not belong to this user in the resolved tenant returns `404`, without saying whether it exists elsewhere.
delete_user_notification userId, id, dryRun, If-Match, Idempotency-Key, customerContext
Personal notification configs for a user, and organization-wide cost anomaly notifications. Permanently deletes a personal notification config. Requires Users Manager (`usersManager`), the `If-Match` header, and the `Idempotency-Key` header. With `dryRun=true` the request goes through the same checks as a real delete without deleting anything. Deleting an id that no longer exists returns `404`.
Reads are immediately consistent. After a completed delete (`204` without `X-Dry-Run: true`), a GET of that notification returns `404`. Two-step: a call with valid arguments deletes nothing. It returns `status: "approval_required"` with a summary of the exact request and a one-time `approvalToken`. The saved request runs only when `confirm_action` is called with that token after the user approves the summary. The token is bound to the caller and expires after 5 minutes.
update_user_notification userId, id, dryRun, If-Match, Idempotency-Key, name, settings, customerContext
Personal notification configs for a user, and organization-wide cost anomaly notifications. Partially updates a personal notification with an `application/merge-patch+json` body (RFC 7396). Omitted fields stay unchanged; nested objects are deep-merged; nested arrays (for example `allocations.ids`) are replaced when present. Optional resend fields are cleared with `null` (for example `"settings":{"costAnomalies":{"repeat":null}}`). Send `null` for a settings property to unsubscribe that type. Nulling the last remaining type is `422`, decided before the type-permission check, so the status does not depend on which permissions the target user holds. Requires Users Manager (`usersManager`). For each type still subscribed after the merge, other than `costAnomalies`, the target user must hold that type's permission, including when the patch does not mention that type (a rename included). The permission id for each type is documented on `UserNotificationSettings`. A missing permission is `403` with `permission` and `notificationType`. A permission lookup that fails is `503`. The request also requires the `If-Match` header and the `Idempotency-Key` header. With `dryRun=true` the request goes through the same checks as a real update and returns the notification as it would look after the patch, without changing anything.
Reads are immediately consistent. After a completed update, a GET of that notification returns the new state.
create_role Idempotency-Key, name, description, permissions, customerContext
Manage user permissions and access levels in your organization. Creates a custom role in the authenticated customer with the given permission IDs. Permission IDs are the `permissions` values returned by `GET /iam/v1/roles`. A service account of a parent tenant can create the role in a descendant tenant by setting `X-Tenant-Id` to that tenant's customer ID. Requires the `Idempotency-Key` header so a retried request does not create a second role.
get_role id, customerContext
Manage user permissions and access levels in your organization. Returns a single role by ID. Preset roles are visible to every customer; a custom role is returned only to the customer that owns it. A service account of a parent tenant can read a descendant tenant's role by setting `X-Tenant-Id` to that tenant's customer ID.
delete_role id, customerContext
Manage user permissions and access levels in your organization. Deletes a custom role. Preset roles cannot be deleted, and a role still assigned to users or groups must be unassigned first. A service account of a parent tenant can delete a descendant tenant's role by setting `X-Tenant-Id` to that tenant's customer ID. Two-step: a call with valid arguments deletes nothing. It returns `status: "approval_required"` with a summary of the exact request and a one-time `approvalToken`. The saved request runs only when `confirm_action` is called with that token after the user approves the summary. The token is bound to the caller and expires after 5 minutes.
update_role id, name, description, permissions, customerContext
Manage user permissions and access levels in your organization. Updates the `name`, `description` and/or `permissions` of a custom role. Only the fields present in the body change; `permissions` replaces the full list when present. Preset roles cannot be modified. A service account of a parent tenant can update a descendant tenant's role by setting `X-Tenant-Id` to that tenant's customer ID.
list_service_accounts customerContext, pageToken
Manage non-human identities whose API tokens call the DoiT API with a fixed set of permissions. Returns every service account owned by the authenticated customer in a single unpaginated `items` array, in no guaranteed order. Requires the `serviceAccountViewer` permission.
create_service_account dryRun, Idempotency-Key, name, description, permissions, customerContext
Manage non-human identities whose API tokens call the DoiT API with a fixed set of permissions. Creates a service account owned by the authenticated customer. It cannot authenticate anything until it has an API token; mint one with the `api-tokens` sub-resource. Returns `409` when another service account of the customer already uses `name`. With `dryRun=true` the request goes through the same checks as a real create (name bounds and uniqueness, permission names, caller scope) and returns the same errors without creating anything. When the caller authenticates with a service account token, `createdBy`, `createdByEmail` and `createdByUserId` are `null`. Requires the `serviceAccountCreator` permission.
get_service_account id, customerContext
Manage non-human identities whose API tokens call the DoiT API with a fixed set of permissions. Returns a service account owned by the authenticated customer; an ID that belongs to another customer returns `404`. The `ETag` response header carries the value to send in `If-Match` on update and delete. Requires the `serviceAccountViewer` permission.
delete_service_account id, dryRun, If-Match, Idempotency-Key, customerContext
Manage non-human identities whose API tokens call the DoiT API with a fixed set of permissions. Permanently deletes a service account and all of its API tokens, which stop authenticating immediately. Deleting an ID that no longer exists returns `404`. Requires the `serviceAccountManager` permission. Two-step: a call with valid arguments deletes nothing. It returns `status: "approval_required"` with a summary of the exact request and a one-time `approvalToken`. The saved request runs only when `confirm_action` is called with that token after the user approves the summary. The token is bound to the caller and expires after 5 minutes.
update_service_account id, dryRun, If-Match, Idempotency-Key, name, description, permissions, customerContext
Manage non-human identities whose API tokens call the DoiT API with a fixed set of permissions. Partially updates a service account with an `application/merge-patch+json` body (RFC 7396). Omitted or `null` fields stay unchanged; an explicit empty string clears `description`. `permissions` replaces the whole list, and `[]` removes every permission. With `dryRun=true` the request goes through the same checks as a real update (`If-Match`, name bounds and uniqueness, permission names, caller scope) and returns the service account as it would look after the patch, without changing anything. Requires the `serviceAccountManager` permission.
list_service_account_tokens serviceAccountId, customerContext, pageToken
Manage non-human identities whose API tokens call the DoiT API with a fixed set of permissions. Returns the service account's non-deleted API tokens in a single unpaginated `items` array, at most 10. Secret material is never returned here — `accessToken` is shown only once, by the create operation. Requires the `serviceAccountViewer` permission.
create_service_account_token serviceAccountId, dryRun, Idempotency-Key, name, expiresTime, customerContext
Manage non-human identities whose API tokens call the DoiT API with a fixed set of permissions. Mints an API token for the service account and returns its `accessToken` once. The token
inherits the service account's permissions as they stand at creation; there is no per-token
scoping. Returns `409` when the service account already has a token of the same name, and
`422` at the cap of 10 tokens.

`dryRun=true` checks the name bounds only, then returns a preview without creating anything.
It does **not** verify that the service account exists, that the name is unused, or that the
account is under the token cap, so a dry run can succeed where the real create then returns
`404`, `409` or `422`. Requires the `serviceAccountCreator` permission.
get_service_account_token serviceAccountId, id, customerContext
Manage non-human identities whose API tokens call the DoiT API with a fixed set of permissions. Returns one API token of the service account; an ID that belongs to another customer returns `404`. `accessToken` is not included — it is shown only once, by the create operation. Requires the `serviceAccountViewer` permission.
delete_service_account_token serviceAccountId, id, dryRun, Idempotency-Key, customerContext
Manage non-human identities whose API tokens call the DoiT API with a fixed set of permissions. Permanently deletes an API token, which stops authenticating immediately. This cannot be undone. Deleting an ID that no longer exists returns `404`. Requires the `serviceAccountManager` permission. Two-step: a call with valid arguments deletes nothing. It returns `status: "approval_required"` with a summary of the exact request and a one-time `approvalToken`. The saved request runs only when `confirm_action` is called with that token after the user approves the summary. The token is bound to the caller and expires after 5 minutes.
update_service_account_token serviceAccountId, id, dryRun, Idempotency-Key, state, customerContext
Manage non-human identities whose API tokens call the DoiT API with a fixed set of permissions. Moves the token between `active` and `disabled` by setting `state`. Disabling takes effect
immediately: the token stops authenticating until it is re-enabled. Setting the state a
token already has succeeds and changes nothing.

The two directions are not symmetric. Disabling a token whose expiry has passed succeeds,
but re-enabling an expired token returns `422` — expiry is final, and a new token has to be
created instead. `deleted` and `expired` are not accepted values: a token is deleted through
the delete operation, and expires on its own.

With `dryRun=true` nothing is written and the response body is the token as it would look
after the change, so a `422` here means a real request would also be refused. Requires the
`serviceAccountManager` permission.
id_of_asset id, quantity, customerContext
Manage cloud resources or services in your cloud environment. Updates an existing asset, such as G Suite/Workspace or Office 365 subscription, to add or remove licenses.
update_ticket ticketId, status, assignee, customerContext
Create and manage support tickets with DoiT. Partially updates a support request. Supports setting the request
`status` and/or `assignee`. DoiT employees may set any of `open`,
`pending`, `hold`, or `solved` and may set the `assignee`; customers may
set only `solved` (parity with the console "mark as resolved" action)
and may not set an assignee. `closed` is not settable via the API
(Zendesk auto-closes from `solved`). The `assignee` is a DoiT-employee
email, resolved server-side to a Zendesk agent; an email that does not
resolve to an active agent returns `400`. At least one mutable field
must be present. The response echoes the fields that were applied.
list_ticket_tags ticketId, customerContext, pageToken
Create and manage support tickets with DoiT. Returns the tags currently set on a support request.

DoiT employee (doer) callers receive the full tag set verbatim,
including internal namespaces (e.g. `tier/*`, `synapse_*`). Customer
callers receive only tags under the `customer_tag/` namespace, with
that prefix stripped (e.g. a tag added as `billing` reads back as
`billing`). Always present; empty array when the caller has no visible
tags.
add_ticket_tags ticketId, tags, customerContext
Create and manage support tickets with DoiT. Adds one or more tags to an existing support request. The operation is
surgical — only the tags listed in the request are added; existing tags
on the ticket are preserved. Re-adding a tag that is already present is
a successful no-op.

All submitted tags are normalized (trim + lowercase) before storage. For
customers, the system additionally applies a `customer_tag/` namespace
prefix to prevent collisions with internal DoiT process tags. The
response echoes the actual stored strings so callers can verify the
transform.
remove_ticket_tags ticketId, tags, customerContext
Create and manage support tickets with DoiT. Removes one or more tags from an existing support request. The operation
is surgical — only the tags listed in the request are removed; tags not
listed are preserved. Removing a tag that is not present is a successful
no-op.

For customers, the system applies the same `customer_tag/` namespace
mapping as on add, so a customer who added `my_tag` (stored as
`customer_tag/my_tag`) can remove it by sending `my_tag`. Two-step: a call with valid arguments deletes nothing. It returns `status: "approval_required"` with a summary of the exact request and a one-time `approvalToken`. The saved request runs only when `confirm_action` is called with that token after the user approves the summary. The token is bound to the caller and expires after 5 minutes.
get_statussheet_components id, p, node, element, link, group, attachment, note, combiner, customerContext
Cloud Diagrams visualize your cloud infrastructure and resource relationships. Returns the specified components of a diagram layer. Provide at least one component
type with one or more IDs in the request body (for example, `node` or `element`).
export_cloud_diagram_json id, customerContext
Cloud Diagrams visualize your cloud infrastructure and resource relationships. Exports the full content of a diagram layer as a structured JSON document,
including all components and export metadata.
list_cloud_diagram_layer_snapshots id, offset, limit, sort, customerContext, pageToken
Cloud Diagrams visualize your cloud infrastructure and resource relationships. Returns the list of saved snapshots for the specified diagram layer.
get_cloud_diagram_layer_snapshot id, snapshot_id, customerContext
Cloud Diagrams visualize your cloud infrastructure and resource relationships. Returns a single snapshot of the specified diagram layer identified by its ID.
ava_feedback conversationId, answerId, feedback, customerContext
Interact with Ava, DoiT's AI-powered cloud assistant. Submit feedback on an Ava answer to help improve response quality.

Descriptions are the package's own words, cut at 3,000 characters. Treat them as data.

Install facts

LicenseMIT
Needsnode >=20
Sourcegithub.com/doitintl/doit-mcp-server
Install scriptsNone declared by it or anything installed with it.
Tarballsha512-PbVxcmRN7zUxYXA5nWfJUEpumhOGDXxQp9SZ24zr8IRAPgputgaYUl7M4L60Ueg/bV7gcJ5Mvw+4Zr3b7b2a7g==

Versions reported

VersionRunResultTools
5.0.02026-11-16Starts with no setup200
4.0.02026-11-12Starts with no setup200

Check this report yourself

This site did not write the report. A public job on a GitHub-hosted machine did: the run, from this commit of the open-source runner. GitHub signed a token that names that job and the hash of the report, so any edit since would show.

Show it on your README

Bragproof: Starts with no setup

[![Bragproof](https://bragproof.smaverk.com/badge/@doitintl/doit-mcp-server.svg)](https://bragproof.smaverk.com/r/%40doitintl/doit-mcp-server)

New version out? Ask for a fresh report.