MCP Tools Reference
Complete catalog of DebugBundle MCP tools with parameters and examples.
This page documents every tool the DebugBundle MCP server exposes. Hosted tools can authenticate
with DEBUGBUNDLE_MEMBER_TOKEN in the MCP server environment, CLI auth state from
~/.debugbundle/auth.json, or an optional per-call bearerToken parameter. bearerToken is
omitted from parameter tables below for brevity.
OpenAI read-only subset
The local stdio server below retains its complete backwards-compatible catalog. The separate OpenAI Plugin 1.0.2 production candidate exposes only these twenty-three read-only tools:
list_projects, list_services, list_incidents, get_incident, get_incident_context, get_bundle, get_reproduction, list_improvements, get_improvement, get_improvement_bundle, get_usage_summary, get_route_metrics, get_journey_patterns, get_device_breakdown, get_referrer_metrics, get_action_metrics, list_funnel_metrics, get_funnel_analysis, get_incident_impact, list_health_checks, get_health_check, list_health_check_results, and list_health_check_daily_rollups.
The OpenAI inputs omit local source selection and credential fields, reject unknown fields, and use OAuth resource scopes including debugbundle:analytics:read. Every operation is read-only and returns an allowlisted structured projection. Mutation, individual analytics journeys, custom analytics dimensions, local filesystem/setup, artifact generation, token, billing, alert, webhook, probe, notification, and member-management tools are excluded. See OpenAI Plugin for its exact auth and privacy boundary.
Project and service consolidation
Ordinary member-authenticated MCP 1.19.0 adds the following tools. The hosted OpenAI read-only subset retains its existing 23-tool contract. See Projects and services for the checked workflow. Mutations require normal per-call approval in OpenClaw.
| Tools | Inputs and behaviour |
|---|---|
list_service_definitions | projectId, optional limit (1–200), cursor; follow next_cursor. |
get_service_definition | projectId, serviceDefinitionId; read logical identity and repository/environments. |
create_service_definition | projectId, service: {name, description?, repository_url?, repository_root?}. |
update_service_definition | projectId, serviceDefinitionId, update; nullable repository fields restore defaults. |
preview_project_merge | request: {target_project_id, source_project_ids, service_mappings}; read-only digest, blockers and suggestions. |
apply_project_merge | request, plan_digest, idempotency_key; owner-checked atomic cutover. |
get_project_merge | Exactly one of operation_id or idempotency_key; recover an uncertain outcome. |
verify_project_merge | operation_id; read-only server verification. |
resume_project_merge | operation_id; recover eligible work. |
cancel_project_merge | operation_id; cancel before commit. |
Existing list_projects accepts optional projectId for one retained view;
existing list_services accepts optional offset for environment pagination.
Existing update_alert, update_webhook and update_github_dispatch_rule accept
optional projectScope: "current_project" to explicitly snapshot the current
whole-project scope; ordinary edits retain their old scope. For capture rules use
update.project_scope: "current_project". activate_probe optionally accepts
sdkProjectId, sdkService and sdkTokenId for explicit deployed credential/name
selection; the default preserves the original signing identity. Existing required
inputs, credential separation and hosted schemas remain unchanged.
Public Status Pages
Public status pages reuse recorded health data and never execute checks. Publication is opt-in, and selected project/check names become public. New checks remain private.
| Tool | Purpose |
|---|---|
get_public_status_page | Read settings and the stable share URL; collaborators receive only a published title/link. |
save_public_status_page | Owner-only save, publish or unpublish using explicit project/check selections. |
list_public_status_page_options | Owner-only paginated projects and checks with the same current owner and account. |
preview_public_status_page | Owner-only preview of the saved safe public projection, including disabled pages. |
Each tool takes projectId. Save also takes settings: {title, enabled, projects: [{project_id, check_ids}]}, with the anchor project first. Options accepts separate cursor or check_project_id/check_cursor pagination. Use the normal member authentication; these tools are excluded from hosted OpenAI and restricted agent-read profiles. See publication setup and the generated schemas.
Setup & Diagnostics
doctor
Run a full connectivity and configuration check.
| Parameter | Type | Description |
|---|---|---|
| (none) | — | Validates local setup, checks API reachability, verifies auth state. |
Returns a diagnostic report with pass/fail status for each check.
validate
Validate your local DebugBundle configuration files.
| Parameter | Type | Description |
|---|---|---|
fix | boolean | Optional. Recreate missing generated setup files when true. |
verify_local
Verify local-mode setup (CLI installed, log directory writable, parser available).
| Parameter | Type | Description |
|---|---|---|
| (none) | — | Checks local-only prerequisites. |
verify_cloud
Verify cloud connectivity (API reachable, token valid, project accessible) or actively prove hosted incident creation.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Cloud project ID. |
service | string | Optional service filter for passive checks and synthetic verification events. |
environment | string | Optional environment filter for passive checks and synthetic verification events. |
maxAgeMinutes | number | Optional lookback window for passive incident checks. |
trigger5xx | boolean | Optional. Send a synthetic 503 request event and verify hosted incident plus bundle visibility. |
trigger4xxStatus | number | Optional. Send a synthetic 400-499 client-error request event and verify hosted incident plus bundle visibility. |
trigger4xxStatus succeeds only when the target project's capture policy promotes that status. It is mutually exclusive with trigger5xx.
verify_app_event
Wait for a real SDK-driven app incident and verify its hosted bundle. This additive
tool preserves the existing verify_cloud contract and uses the same verifier and
result shape as debugbundle verify cloud --expect-app-event --json.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required project ID. |
service | string | Service to match. |
traceId | string | Trace correlation hint to match in the bundle. |
requestId | string | Request correlation hint to match in the bundle. |
environment | string | Optional environment; defaults to production. |
maxAgeMinutes | number | Optional verification lookback window; defaults to 15. |
authFilePath | string | Optional path to local member authentication. |
Provide at least one nonblank service, traceId, or requestId. Synthetic trigger
options and unknown fields are rejected. The tool never generates synthetic events;
it reads existing incident and bundle evidence. It is available in ordinary and
local-auth stdio MCP and as debugbundle_verify_app_event in OpenClaw, not the
restricted hosted OpenAI connection or agent-read profile.
smoke
Run a quick smoke test that sends a test event and verifies end-to-end processing.
| Parameter | Type | Description |
|---|---|---|
project_id | string | Optional. Target project for the test event. |
Incident Retrieval
list_incidents
List incidents for the active project.
| Parameter | Type | Description |
|---|---|---|
source | string | Filter by source: local, cloud, or all. Default: all. |
projectId | string | Optional project filter. |
environment | string | Optional environment filter. |
service | string | Filter by service name. |
status | string | Filter: active, open, resolved, regressed, or all. Default: active; active means open or regressed. |
severity | string | Filter by severity: low, medium, high, or critical. |
firstSeenAfter | string | Optional ISO-8601 lower bound for first_seen_at. |
attentionAfter | string | Optional ISO-8601 lower bound for incidents opened or regressed after this time. |
cursor | string | Optional pagination cursor. |
limit | number | Max results. Default: 20. |
get_incident
Get full details of a specific incident.
| Parameter | Type | Description |
|---|---|---|
incident_id | string | Required. The incident ID. |
resolve_incident
Mark an incident as resolved.
| Parameter | Type | Description |
|---|---|---|
incidentId | string | Required. The incident ID. |
resolve_incidents
Mark incidents as resolved in bulk.
| Parameter | Type | Description |
|---|---|---|
incidentIds | string[] | Required. Incident IDs to resolve. |
reopen_incident
Reopen a previously resolved incident.
| Parameter | Type | Description |
|---|---|---|
incidentId | string | Required. The incident ID. |
reopen_incidents
Reopen incidents in bulk.
| Parameter | Type | Description |
|---|---|---|
incidentIds | string[] | Required. Incident IDs to reopen. |
get_bundle
Retrieve the debug bundle for an incident.
| Parameter | Type | Description |
|---|---|---|
incident_id | string | Required. The incident ID. |
format | string | Output format: json (default) or summary. |
get_reproduction
Get the reproduction artifact for an incident.
| Parameter | Type | Description |
|---|---|---|
incident_id | string | Required. The incident ID. |
Returns reproduction commands (cURL, HTTPie) and JSON specifications when available.
get_logs
Retrieve raw log entries associated with an incident.
| Parameter | Type | Description |
|---|---|---|
incident_id | string | Required. The incident ID. |
limit | number | Max log entries. Default: 100. |
AnalyticsBundle
Project flow tools are list_analytics_flows, save_analytics_flow,
archive_analytics_flow, and get_analytics_flow_report. All accept projectId and member
authentication; save takes definition, archive/report take flowKey, and report accepts
window (7d, 30d, 90d, default 30d). Save/archive require owner/admin. The same
public definitions and partial-coverage reports are available in the API and CLI, for any
customer project. These tools are outside the separate hosted OpenAI read-only catalog.
Analytics tools require member authentication. Project tokens are write-only SDK
credentials and cannot call these tools. Metric/detail operations remain project-scoped;
only opportunity and bundle lists may omit projectId for an accessible organization-wide
inventory.
Common metric options are projectId, last, from, to, granularity, service,
environment, and limit where supported.
get_usage_summary
Return sessions, page views, active/new/returning visitors, exits, conversions, and standard aggregate breakdowns for a project.
get_route_metrics
Return aggregate route pageviews, unique sessions, entrances, exits, bounces, and incident-linked sessions.
get_device_breakdown
Return device type, browser, OS, and language aggregates.
get_referrer_metrics
Return referrer-domain and bounded UTM source/medium/campaign aggregates.
get_action_metrics
Return semantic action, conversion, and fixed journey-marker aggregates.
list_funnel_metrics
List funnel entered/completed/drop-off/conversion summaries for a project.
get_funnel_analysis
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project that owns the funnel evidence. |
funnelKey | string | Required. Funnel key to analyze. |
Returns deterministic ordered step metrics in addition to the aggregate funnel summary.
get_journey_patterns
Return aggregate route transitions, unique sessions, transition share, and bounded retained sample references.
list_analytics_journey_samples
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project scope. |
service / environment | string | Optional bounded scope. |
tag | string | Optional derived analysis tag. |
cursor / limit | string / number | Optional pagination. |
get_analytics_journey_sample
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project scope. |
sampleId | string | Required. Retained sample identifier. |
Returns a redacted structured timeline, not a raw visit or video replay.
get_incident_impact
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project scope. |
incidentId | string | Required. Accessible incident to correlate. |
Returns exact correlation-backed affected sessions, routes, funnels, segments, journey references, conversion availability, and current incident-impact bundle state.
list_analytics_opportunities
projectId is optional. Omit it only for an accessible cross-project inventory. Optional
filters include status, kind, service, environment, severity, bundle status, date bounds,
cursor, and limit.
get_analytics_opportunity
Requires projectId and opportunityId; returns one aggregate-only opportunity and its
current generated-bundle state.
list_analytics_bundles
projectId is optional for cross-project inventory. Optional filters include generation
status, analysis kind, service, environment, date bounds, cursor, and limit.
generate_analytics_bundle
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project scope. |
analysisKind | string | Required. Bounded analysis kind. |
last or from/to | string | Bounded analysis window. |
funnel, route, incidentId, deployId | string | Optional kind-specific focus. |
filters | object | Optional safe service/environment scope. |
Returns pending/failed state or a completed deterministic artifact. It does not generate one bundle per visit.
get_analytics_bundle
Requires projectId and bundleGenerationId; returns the existing generation status or completed
artifact without requesting another generation.
get_analytics_settings
Requires projectId. Owners/admins receive manage access mode; members receive preview.
update_analytics_settings
Requires owner/admin access plus projectId and an update object. Settings include enabled
state, privacy, consent, capture toggles, sampling, retention, saved-funnel cap, and Team
controlled custom dimensions.
list_saved_analytics_funnels
List active definitions for projectId. Members may read definitions.
create_saved_analytics_funnel
Requires owner/admin access, projectId, funnelKey, displayName, and 2-20 ordered unique
steps containing step_key and display_name.
update_saved_analytics_funnel
Requires owner/admin access, projectId, funnelKey, and at least one of displayName or
ordered steps.
archive_saved_analytics_funnel
Requires owner/admin access, projectId, and funnelKey. Archival removes the active
definition but retains historical aggregate evidence.
Token Management
list_project_tokens
List project tokens (SDK/ingestion tokens).
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project that owns the tokens. |
limit | number | Optional maximum number of tokens to return. |
create_project_token
Create a new project token for SDK ingestion.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project that owns the token. |
label | string | Required. Human-readable token label. |
allowedOrigins | string[] | Optional origin allowlist for direct/static browser ingestion. |
Returns the token plaintext once — store it immediately. Use allowedOrigins only for public write-only project tokens intentionally embedded in static browser apps. Allowlisted tokens reject requests without a matching Origin, so do not reuse them for server SDKs or relay forwarding. This reduces normal browser-origin abuse but is not a secret boundary.
revoke_project_token
Revoke a project token.
| Parameter | Type | Description |
|---|---|---|
token_id | string | Required. The token ID to revoke. |
list_member_tokens
List member tokens (CLI/API/MCP tokens).
| Parameter | Type | Description |
|---|
create_member_token
Create a new member token.
| Parameter | Type | Description |
|---|---|---|
label | string | Required. Human-readable token label. |
Returns the token plaintext once — store it immediately.
revoke_member_token
Revoke a member token.
| Parameter | Type | Description |
|---|---|---|
token_id | string | Required. The token ID to revoke. |
list_agent_tokens
List metadata for restricted agent credentials on one project. Requires project owner/admin access through the ordinary member-authenticated MCP connection; token plaintext is never returned by this list.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project to inspect. |
create_agent_token
Create a single-project, read-only agent credential. Issuance is disabled by default on the API until an operator completes the mixed-fleet migration gate. The plaintext credential is returned once; store it in a protected environment or the restricted CLI profile.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project the agent may read. |
label | string | Required. Human-readable credential label. |
expiresAt | string | Optional ISO timestamp, at most 90 days in the future. |
revoke_agent_token
Revoke a restricted agent credential for one project. Existing agent reads fail after revocation.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project that owns it. |
tokenId | string | Required. Credential ID to revoke. |
Webhook Management
list_webhooks
List configured webhook endpoints.
| Parameter | Type | Description |
|---|---|---|
project_id | string | Optional. Defaults to active project. |
create_webhook
Create a new webhook endpoint.
| Parameter | Type | Description |
|---|---|---|
project_id | string | Optional. Defaults to active project. |
url | string | Required. The webhook delivery URL. |
events | string[] | Event types to subscribe to. Default: all events. |
secret | string | Optional signing secret for HMAC verification. |
update_webhook
Update an existing webhook.
| Parameter | Type | Description |
|---|---|---|
webhook_id | string | Required. The webhook ID. |
url | string | New delivery URL. |
events | string[] | Updated event subscriptions. |
is_active | boolean | Enable or disable the webhook. |
delete_webhook
Delete a webhook endpoint.
| Parameter | Type | Description |
|---|---|---|
webhook_id | string | Required. The webhook ID. |
test_webhook
Send a test delivery to a webhook endpoint.
| Parameter | Type | Description |
|---|---|---|
webhook_id | string | Required. The webhook ID. |
test_webhook_event
Queue a signed synthetic test for any of the eight supported webhook events. This
additive tool retains the legacy test_webhook verification-only schema. The test
uses the existing synthetic envelope and does not create a real incident or bundle.
| Parameter | Type | Required | Description |
|---|---|---|---|
projectId | string | Yes | Project owning the endpoint. |
webhookId | string | Yes | Endpoint to test. |
eventType | string | Yes | Any supported lifecycle or verification webhook event. |
Requires member authentication and endpoint creator or owner/admin authority. The hosted OAuth and restricted agent-read profiles do not expose this mutation.
list_webhook_deliveries
List recent deliveries for a webhook.
| Parameter | Type | Description |
|---|---|---|
webhook_id | string | Required. The webhook ID. |
limit | number | Max results. Default: 20. |
retry_webhook_delivery
Retry a failed webhook delivery.
| Parameter | Type | Description |
|---|---|---|
delivery_id | string | Required. The delivery ID. |
Alert Management
list_alerts
List configured alert rules.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project that owns the alert rules. |
limit | number | Optional maximum number of alerts to return. |
list_alert_groups
List direct alert deliveries and email digests for one project. Results contain safe status, channel, timestamps, and incident counts without destination secrets or raw event payloads.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project UUID. |
limit | number | Optional page size, 1–100; default 50. |
cursor | string | Opaque cursor returned by the preceding group page. |
get_alert_group
Inspect one group's safe metadata and a page of member incident IDs. Use its returned cursor for the next member page.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project UUID. |
kind | string | Required. direct or email_digest. |
groupId | string | Required. Group UUID. |
limit | number | Optional member page size, 1–100; default 50. |
cursor | string | Opaque cursor from the preceding member page. |
create_alert
Create a new alert rule.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project that owns the alert rule. |
serviceId | string | Optional service scope. |
channel | string | Required. Alert channel: email, slack, discord, webhook. |
conditionType | string | Required. Trigger: new_incident, incident_regressed, error_spike, severity_threshold, regression_after_deploy. |
severityMin | string | Minimum severity filter. |
severityLifecycleScope | string | For severity_threshold only: new_incident, incident_regressed, or both. Defaults to both; non-threshold rules return null. |
cooldownSeconds | number | Optional notification cooldown in seconds. Defaults to 0, accepts 0 to disable suppression, max 604800 (7 days). |
config | object | Required. Channel-specific configuration. |
isEnabled | boolean | Optional enabled state. Defaults to true. |
For conditionType: "severity_threshold", severityLifecycleScope: "both" notifies
for matching new incidents and matching regressions. Non-threshold conditions ignore
this field because their condition already defines the lifecycle event.
severityMin: "high" does not match low-confidence opaque browser-native captures by
default: opaque window_error infers low, and opaque resource_error infers
medium.
For channel: "webhook", the updated MCP client requests a signed rule. Its creation
response reveals alert.signing_secret once. Store it in the receiver and verify the
X-DebugBundle-Signature: sha256=<hex> HMAC over the exact request body. Older rules
also receive a signature header after a private key is provisioned; rotate a rule
to reveal a new key for receiver verification.
update_alert
Update an existing alert rule.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project that owns the alert rule. |
alertId | string | Required. The alert ID. |
serviceId | string | null | Update or clear service scope. |
channel | string | Updated channel type. |
conditionType | string | Updated trigger condition. |
severityMin | string | null | Updated or cleared severity filter. |
severityLifecycleScope | string | null | Updated severity-threshold lifecycle scope; use null to reset to the default both. |
cooldownSeconds | number | Updated notification cooldown in seconds. Use 0 to disable suppression. |
isEnabled | boolean | Enable or disable the alert. |
config | object | Updated channel configuration. |
rotate_alert_webhook_secret
Rotate a custom alert-webhook signing secret. The new alert.signing_secret appears
only in this response. Update the receiver before requiring the new signature. The
existing update_alert tool schema remains unchanged.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project that owns the rule. |
alertId | string | Required. Webhook alert rule ID. |
delete_alert
Delete an alert rule.
| Parameter | Type | Description |
|---|---|---|
alert_id | string | Required. The alert ID. |
Slack Destinations
list_slack_destinations
List connected Slack destinations for a project organization.
Saved destinations remain readable after a downgrade to Free so agents can inspect the paused setup before the owner upgrades again.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project whose organization owns the destinations. |
get_slack_connect_url
Return a browser URL for Slack OAuth and channel selection.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project to return to after connection. |
returnTo | string | Optional app-relative return path. |
test_slack_destination
Send a test message to a connected Slack destination.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project whose organization owns the destination. |
destinationId | string | Required. Slack destination ID. |
delete_slack_destination
Delete a connected Slack destination when no alert rule or weekly report still references it. This cleanup action remains available after downgrade.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project whose organization owns the destination. |
destinationId | string | Required. Slack destination ID. |
Weekly Reports
list_weekly_reports
List available weekly reports.
| Parameter | Type | Description |
|---|---|---|
project_id | string | Optional. Defaults to active project. |
limit | number | Max results. Default: 10. |
get_weekly_report
Get a specific weekly report with full analysis.
| Parameter | Type | Description |
|---|---|---|
report_id | string | Required. The report ID. |
generate_weekly_report
Trigger generation of a new weekly report.
| Parameter | Type | Description |
|---|---|---|
project_id | string | Optional. Defaults to active project. |
configure_weekly_reports
Configure weekly report settings.
| Parameter | Type | Description |
|---|---|---|
project_id | string | Optional. Defaults to active project. |
is_enabled | boolean | Enable or disable weekly reports. |
recipients | string[] | Email addresses to receive reports. |
Services
list_services
List all services discovered from ingested events.
| Parameter | Type | Description |
|---|---|---|
project_id | string | Optional. Defaults to active project. |
Projects
list_projects
List projects in the organization.
| Parameter | Type | Description |
|---|---|---|
limit | number | Max results to return. |
create_project
Create a new project.
| Parameter | Type | Description |
|---|---|---|
name | string | Required. Human-readable project name. |
slug | string | Required. URL-friendly identifier (lowercase, hyphens). |
environmentDefault | string | Editable project default environment used in setup snippets and project metadata (for example production). |
update_project
Update an existing project.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. The project ID. |
name | string | New project name. |
slug | string | New URL-friendly identifier. |
environmentDefault | string | New editable project default environment. |
delete_project
Delete a project and all associated data.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. The project ID. |
Health Checks
list_health_checks
List hosted health checks for a project, including current plan limits.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project that owns the hosted health checks. |
limit | number | Optional maximum number of checks to return. |
get_health_check
Get one hosted health check by id.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project that owns the check. |
checkId | string | Required. Health check ID. |
create_health_check
Create a hosted health check for a public HTTP endpoint.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project that owns the check. |
name | string | Required. Human-readable check name. |
url | string | Required. External http or https URL. |
intervalSeconds | number | Required. Execution interval in seconds, subject to plan minimums. |
method | string | Optional GET or HEAD. Defaults to GET. |
expectedStatusMin | number | Optional lowest successful status. Defaults to 200. |
expectedStatusMax | number | Optional highest successful status. Defaults to 399. |
timeoutMs | number | Optional request timeout in milliseconds. |
failureThreshold | number | Optional consecutive failures required to open or regress an incident. Omit for the plan recommendation (2 on Team, 3 otherwise); set 1 deliberately for an especially critical endpoint. |
recoveryThreshold | number | Optional consecutive successes required to recover. |
environment | string | Optional environment label. |
serviceName | string | null | Optional service label; pass null to clear on update. |
enabled | boolean | Optional enabled state. Disabled checks remain visible but do not execute. |
update_health_check
Update a hosted health check. Use this to change thresholds, status expectations, labels, or enabled state.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project that owns the check. |
checkId | string | Required. Health check ID. |
name | string | Optional new name. |
url | string | Optional new target URL. |
intervalSeconds | number | Optional execution interval in seconds. |
method | string | Optional GET or HEAD. |
expectedStatusMin | number | Optional lowest successful status. |
expectedStatusMax | number | Optional highest successful status. |
timeoutMs | number | Optional request timeout in milliseconds. |
failureThreshold | number | Optional consecutive failures required to open or regress an incident. |
recoveryThreshold | number | Optional consecutive successes required to recover. |
environment | string | Optional environment label. |
serviceName | string | null | Optional service label; pass null to clear. |
enabled | boolean | Optional enabled state. |
delete_health_check
Delete a hosted health check.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project that owns the check. |
checkId | string | Required. Health check ID. |
test_health_check
Run a side-effect-free target test using the same validation and guarded request execution as saved checks. This does not create incidents, retained result rows, or rollups.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project used for authorization and plan context. |
url | string | Required. External http or https URL to test. |
method | string | Optional GET or HEAD. Defaults to GET. |
expectedStatusMin | number | Optional lowest successful status. |
expectedStatusMax | number | Optional highest successful status. |
timeoutMs | number | Optional request timeout in milliseconds. |
list_health_check_results
List recent execution results for one hosted health check.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project that owns the check. |
checkId | string | Required. Health check ID. |
limit | number | Optional maximum number of results to return. |
list_health_check_daily_rollups
List retained per-day status history for one hosted health check.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. Project that owns the check. |
checkId | string | Required. Health check ID. |
limit | number | Optional maximum number of rollups to return. |
Probes
activate_probe
Activate a remote debug probe on a project.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. The project ID. |
labelPattern | string | Required. Glob pattern matching incident labels (e.g., auth.*). |
service | string | Scope to a specific service name. |
environment | string | Scope to a specific environment. |
ttlSeconds | number | Probe lifetime in seconds. Default: 3600. Max: 86400. |
triggerTtlSeconds | number | Trigger token validity in seconds. Default: 300. |
Returns the probe object including a one-time trigger_token.
list_active_probes
List active probes for a project.
Saved remote probe activations remain readable after a downgrade to Free so agents can inspect the paused setup before the owner upgrades again.
Probe deactivation remains available as cleanup after downgrade; new remote activation requires the project to return to Solo or Team.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. The project ID. |
deactivate_probe
Deactivate a probe.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. The project ID. |
activationId | string | Required. The probe activation ID. |
Capture Policy
get_capture_policy
Get the resolved capture policy for a project.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. The project ID. |
update_capture_policy
Update the capture policy for a project.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. The project ID. |
update | object | Required. Policy update with optional fields: preset, capture_logs, capture_request_events, capture_breadcrumbs, capture_probe_events, immediate_client_error_statuses, and immediate_client_error_path_rules. Path rules use status_code, path_pattern, and optional methods to promote only matching 4xx combinations, such as 404 under /checkout/*. |
Capture Rules
list_capture_rules
List project capture rules.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. The project ID. |
create_capture_rule
Create a project capture rule.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. The project ID. |
rule | object | Required. Rule payload with action, matcher, and optional sample fields. |
update_capture_rule
Update a project capture rule.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. The project ID. |
ruleId | string | Required. The capture rule ID. |
update | object | Required. Partial rule update. |
delete_capture_rule
Delete a project capture rule.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. The project ID. |
ruleId | string | Required. The capture rule ID. |
suggest_capture_rules_from_incident
Suggest capture rules from an incident bundle without changing project state.
| Parameter | Type | Description |
|---|---|---|
incidentId | string | Required. The incident ID. |
create_capture_rule_from_incident_suggestion
Create a project capture rule from one of the incident suggestions.
| Parameter | Type | Description |
|---|---|---|
incidentId | string | Required. The incident ID. |
suggestionId | string | Required. The suggestion ID returned by suggest_capture_rules_from_incident. |
name | string | Optional override for the rule name. |
description | string | Optional override for the rule description. |
enabled | boolean | Optional initial enabled state. |
expiresAt | string | Optional ISO timestamp for automatic expiry. |
Billing
billing_plan_change
Available in core 1.19.0 and published MCP package 1.18.0.
Inspect, preview, schedule or cancel a Team-to-Solo change at the next renewal on the existing subscription. Owner access is required. Purchased extra capacity is retained at Solo pricing; Team stays active until the effective date.
| Parameter | Type | Description |
|---|---|---|
action | string | Required: status, preview, schedule or cancel. |
confirmationToken | string | Required for schedule: the exact signed token returned by preview. |
scheduleId | string | Required for cancel: the exact pending schedule from status. |
Present the preview's effective date, estimated monthly total, tax, retained capacity and loss of Team features, then obtain explicit confirmation before scheduling. Nothing is charged or refunded immediately. Retry an uncertain confirmed request with its original token; do not invent a new change or create another subscription. Cancel releases the pending change and keeps Team. This tool is part of the ordinary member-authenticated server, not the restricted read-only connection.
get_billing_summary
Get the billing summary for the organization.
| Parameter | Type | Description |
|---|---|---|
| (none beyond auth) | — | Returns plan, billing state, trial state, active-project counts, capacity units, and scheduled reduction info. |
start_trial
Start an eligible no-card trial for the organization.
| Parameter | Type | Description |
|---|---|---|
targetPlan | string | Required. solo or team. |
increase_capacity
Add extra capacity units with prorated billing.
| Parameter | Type | Description |
|---|---|---|
targetAdditionalCapacityUnits | number | Required. Target number of additional capacity units. Active no-card trials receive trial_conversion_required. |
schedule_capacity_reduction
Schedule a capacity reduction at the end of the billing cycle.
| Parameter | Type | Description |
|---|---|---|
targetAdditionalCapacityUnits | number | Required. Target number of additional capacity units after reduction. Active no-card trials receive trial_conversion_required. |
cancel_capacity_reduction
Cancel a scheduled capacity reduction.
| Parameter | Type | Description |
|---|---|---|
| (none beyond auth) | — | Cancels the pending reduction, keeping current capacity count. Active no-card trials receive trial_conversion_required. |
Members
list_project_members
List members for a project.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. The project ID. |
list_project_member_invites
List pending invites for a project.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. The project ID. |
invite_project_member
Invite a collaborator to a project.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. The project ID. |
email | string | Required. Email address for the invitation. |
role | string | Required. Member role (e.g., admin, member). |
cancel_project_member_invite
Cancel a pending invitation.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. The project ID. |
inviteId | string | Required. The invite ID to cancel. |
update_project_member_role
Update the role of a project member.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. The project ID. |
userId | string | Required. The user ID. |
role | string | Required. The new role. |
remove_project_member
Remove a member from a project.
| Parameter | Type | Description |
|---|---|---|
projectId | string | Required. The project ID. |
userId | string | Required. The user ID to remove. |
Analysis
analyze
Run deep analysis on an incident, combining bundle data, logs, and reproduction artifacts.
| Parameter | Type | Description |
|---|---|---|
incident_id | string | Required. The incident ID. |
Returns a structured analysis including root cause hypothesis, affected components, severity assessment, and suggested fixes.
GitHub installation cleanup
disconnect_github_installation
Disconnect the installation for the active organization and remove repository assignments across all its projects. Requires the organization owner; cleanup remains available after downgrade. The App stays installed on GitHub. Install/reconnect uses the project GitHub dashboard browser flow to preserve its signed state cookie.
| Parameter | Type | Required | Description |
|---|---|---|---|
bearerToken | string | Yes | Member token for the active organization owner |
Next Steps
- MCP Server — Overview and authentication
- MCP Workflows — Practical agent patterns
Notification preferences
The ordinary MCP catalog exposes get_notification_preferences,
update_notification_preferences, pause_notifications, and
resume_notifications. Common input is bearerToken, optional
audience: "personal" | "shared" (default personal) and optional project UUID
projectId. Local-auth mode supplies credentials through the existing profile.
Update takes update: { channels: { email: "daily" }, daily: { time: "09:00", timezone: "Europe/Ljubljana" } }. Pause takes pause: { channels: ["email"], until: null, allow_health_checks: false }; use a future UTC ISO timestamp instead
of null for a timed pause. Resume takes resume: { channels: ["email"] }.
Channel modes and authorization match the API.
These management operations are excluded from the hosted read-only OpenAI and restricted-agent profiles. OpenClaw treats new operations as optional tools.