¿En qué se diferencia de conectar JFrog MCP directamente?

Conectado directamente, un agente ve todas las herramientas que expone el servidor y guarda él mismo la credencial. A través de TrustGate tú eliges qué herramientas recibe cada agente, y cada llamada queda registrada a nombre del agente que la hizo.

¿Qué pasa si cambia la API de JFrog MCP?

Seguimos el esquema de JFrog MCP, así que el catálogo de herramientas se mantiene al día. Las herramientas que ya habías permitido siguen funcionando, y cualquier novedad queda denegada hasta que la actives.

¿Puedo dar herramientas de JFrog distintas a cada agente?

Sí. El acceso se define por agente: uno puede recibir unas pocas herramientas de lectura y otro el conjunto completo. Concedes el MCP entero o herramientas sueltas dentro de él.

¿Dónde veo lo que han hecho mis agentes en JFrog?

Cada llamada queda registrada a nombre del agente que la hizo, junto con la herramienta que usó. La actividad puede enviarse a tu propio sistema de monitorización.

¿Tengo que alojar yo el servidor de JFrog MCP?

No. TrustGate se conecta al servidor de JFrog MCP y se sitúa delante de él. Tu agente solo habla con tu URL de TrustGate, nunca directamente con JFrog.

Conecta JFrog MCP a tus agentes de IA, de forma segura.

Q: ¿Es seguro JFrog MCP?

El servidor expone 114 herramientas y algunas pueden escribir en tu cuenta de JFrog. Detrás de TrustGate, un agente solo ve las herramientas que tú permites, las peticiones y respuestas se inspeccionan y cada llamada queda registrada. Todo lo que esté fuera de la lista permitida se deniega antes de llegar a JFrog.

Un único endpoint seguro para cada llamada a JFrog que hacen tus agentes de IA: acotada por usuario, inspeccionada y registrada.

Empieza gratis

Solicita una demo

Prueba JFrog MCP ahora

Lo que pueden pedir tus agentes una vez conectado JFrog MCP.

Elige las herramientas que pueden usar tus agentes de IA

114 herramientas del esquema de JFrog MCP.

Mostrando 114 herramientas

access_tokens_createhigh riskCreates a new access token, optionally non-expirable and scoped to a project. Beta.
access_tokens_listreadLists access tokens with filters for username, description, token ID and refreshable status. Beta.
access_oidc_create_configurationhigh riskCreates an OIDC identity provider configuration, defining a new trusted token issuer.
access_projects_createwriteCreates a new Access project with its admin privileges and storage quota.
access_projects_getreadGets the details of an Access project.
access_projects_updatewriteUpdates an Access project's name, description, privileges or quota.
access_projects_create_rolewriteCreates a custom project role with its environments and allowed actions.
access_projects_update_rolewriteUpdates a project role. Only CUSTOM roles can be updated.
access_projects_get_userreadGets a user's membership in a project.
access_projects_upsert_userhigh riskAdds or updates a user in a project. The roles given replace all existing roles.
access_projects_list_groupsreadLists the groups assigned to a project.
access_projects_upsert_groupwriteAdds or updates a group in a project. The roles given replace all existing roles.
access_projects_get_repositoryreadGets a repository's project status.
access_projects_move_repository_to_projectwriteMoves a repository into a project, optionally forcing past validation checks.
access_projects_share_repository_with_allhigh riskShares a repository with every project on the platform, optionally read-only.
access_projects_listreadLists all JFrog projects.
access_roles_create_global_rolehigh riskCreates a custom global role with platform-wide permissions. Access 7.77 and later.
access_roles_update_global_rolewriteUpdates an existing custom global role. Access 7.77 and later.
access_stages_list_global_stagesreadLists the global stages configured in the platform, filterable by project, scope and category.
access_users_listreadLists platform users, filterable by status, username, admin-only or project.
access_users_getreadGets the details of a user by username.
access_groups_listreadLists all groups in the platform.
access_groups_getreadGets the details of a group by name.
artifactory_repositories_createwriteCreates a local, remote, virtual or federated repository.
artifactory_repositories_getreadRetrieves the configuration of a single repository by key.
artifactory_repositories_listreadLists repositories, filterable by type, package type and project.
artifactory_repositories_updatewriteUpdates an existing repository's configuration, including Xray indexing and inclusion patterns.
artifactory_packages_get_versionsreadLists the versions of a package stored in Artifactory with summary information. Artifactory 7.146.0 and later.
artifactory_packages_get_version_repositoriesreadFinds which repositories hold a specific package version. Artifactory 7.146.0 and later.
artifactory_packages_get_install_instructionsreadGets the native client commands for installing a specific package version. Artifactory 7.133.3 and later.
artifactory_builds_list_buildsreadLists builds using AQL, returning names, numbers and timestamps. The query must use builds.find() and include a limit.
artifactory_builds_list_build_runsreadLists all runs of a named build, newest first.
artifactory_builds_get_inforeadGets the full build info for a run: metadata, modules, artifacts, dependencies and CI/VCS information, optionally diffed against an older build.
artifactory_federation_get_statusreadGets the health status of federated repositories.
artifactory_federation_get_conflictsreadFinds federated repositories with configuration conflicts.
artifactory_federation_update_repo_configwriteSynchronizes a federated repository's configuration to its members.
artifactory_storage_artifact_inforeadRetrieves storage information about a file or folder, with optional deep listing, download statistics, properties and permissions. Artifactory 7.135 and later.
xray_artifact_get_summaryreadGets the Xray security summary of one or more artifacts by path or checksum.
xray_artifact_get_violationsreadGets Xray policy violation counts for one artifact across one or more versions.
xray_artifact_security_statusreadGets the Xray indexing and download-block status of a single artifact. Xray 3.147.2 and later.
xray_sbom_search_impacted_resourcesreadSearches impacted resources by vulnerability, package, package version or SaaS service.
catalog_vulnerabilities_getreadGets vulnerability information for a CVE, including affected packages and versions.
catalog_packages_versions_vulnerabilitiesreadLists the known vulnerabilities for a specific package version.
catalog_packages_list_versionsreadLists versions of a publicly available package, filterable by vulnerability status.
catalog_packages_getreadGets public information about a software package.
jfs_curation_check_remote_package_compliancereadChecks Curation compliance for a third-party package before installing or recommending it. Xray 3.147.2 and later.
jfs_curation_query_audit_eventsreadSearches blocked Curation package audit events. Covers policy blocks only, not approved or dry-run events.
jfs_curation_query_cvs_audit_eventsreadSearches Compliant Version Selection audit events: which versions were approved or blocked and the policies involved. Xray 3.153.0 and later.
jfs_curation_get_cvs_audit_version_detailreadGets the full policy explanation and remediation for one package version evaluated by CVS.
jfs_curation_get_audit_event_policiesreadGets the policies, explanation and remediation behind one download-time Curation block.
jfs_curation_create_waiver_requestwriteCreates a waiver request for a package blocked by a Curation policy. Xray 3.148.0 and later.
jfs_curation_query_waiver_requestsreadSearches and checks the status of Curation waiver requests.
jfs_curation_decide_waiver_requesthigh riskApproves or rejects a pending waiver request, unblocking a package the policy rejected. Xray 3.149.0 and later.
distribution_release_bundles_distributehigh riskStarts a Release Bundle v2 distribution to one or more edge nodes. Can bypass the Xray scan block and auto-create missing repositories.
distribution_release_bundles_abortwriteAborts an in-progress distribution. Does not roll back already-transferred files.
distribution_release_bundles_get_path_mappingreadGets the applied path mapping and ATW flag for a release bundle version.
distribution_trackers_listreadLists the distribution trackers for a release bundle version.
distribution_trackers_get_by_idreadGets per-edge progress for a specific distribution tracker.
distribution_signing_keys_get_for_bundlereadGets the signing keys associated with a release bundle version.
distribution_signing_keys_get_publicreadRetrieves the public part of a Distribution signing key as PEM.
distribution_signing_keys_propagatehigh riskPushes a public signing key to every edge node.
distribution_edges_list_for_bundlereadLists the edge nodes eligible for a release bundle distribution or deletion.
aicatalog_skills_listreadLists the skills published in the JFrog AI Catalog.
aicatalog_agent_plugins_listreadLists the agent plugins published in the JFrog AI Catalog.
aicatalog_mcps_listreadLists the MCP servers the AI Catalog has approved for a project.
aicatalog_mcps_allowhigh riskApproves an MCP server for use in a project, widening what every agent there can reach.
aicatalog_mcps_restrictwriteRemoves an MCP server's approval for a project.
grid_jpds_listreadLists the JPDs known to Mission Control for Grid topology membership.
grid_topology_listreadLists all Grid topologies with their member sites and connection status.
grid_topology_getreadGets one Grid topology by key, including member sites and connection status.
grid_monitoring_summarize_entitiesreadGets aggregated entity-sync counts per type for a Grid topology.
grid_monitoring_list_entitiesreadLists entity IDs and sync status for one entity type in a Grid topology.
grid_monitoring_get_entityreadGets per-target sync detail for one entity, including failure descriptions.
worker_list_allreadLists the JFrog Workers on the platform.
worker_get_specificreadRetrieves the full details of a single Worker by key.
worker_executehigh riskExecutes a GENERIC_EVENT Worker by key with a custom JSON payload. Beta.
worker_list_actionsreadLists the Worker trigger actions available.
worker_get_instructions_for_code_generationreadReturns a code template, type definitions, PlatformContext typings and best-practice instructions for implementing Worker code.
worker_createhigh riskCreates a Worker from full TS/JS source, with its secrets and properties. The Worker is created disabled.
worker_updatehigh riskUpdates an existing Worker, including its code, secrets and enabled state. Beta.
worker_get_execution_historyreadRetrieves the execution history for Workers, optionally including test runs.
onemodel_graphql_query_introspectreadDescribes the OneModel schema for one or more entities so a query can be built. Beta, OneModel 1.109.0 and later.
onemodel_graphql_query_executereadRuns a GraphQL query against the OneModel unified data endpoint. Beta.
event_domains_listreadLists the webhook domain and event type catalog on the Event service.
event_subscriptions_listreadLists the webhook subscriptions the caller can read.
event_subscriptions_getreadGets one webhook subscription by key.
event_subscriptions_createhigh riskCreates a webhook subscription that delivers platform events to an external handler URL.
event_subscriptions_updatewriteUpdates a webhook subscription as a full replace. Beta.
event_troubleshooting_statusreadChecks whether webhook delivery recording and the troubleshooting data API are available.
event_troubleshooting_listreadLists the latest webhook delivery attempt per subscription.
event_troubleshooting_countreadCounts webhook delivery records and errors since a troubleshooting stream start ID.
evidence_config_getreadRetrieves the Evidence service configuration exposed to clients. Evidence 7.277 and later.
evidence_config_get_categoriesreadRetrieves the full Evidence category configuration.
evidence_config_update_categorieswriteReplaces the Evidence categories configuration. Beta.
evidence_categories_listreadLists the predicate categories supported by JFrog Evidence.
evidence_categories_get_predicate_typesreadRetrieves all predicate type URIs belonging to a specific Evidence category.
evidence_records_searchreadSearches evidence records by repository, subject path, subject name or SHA-256.
evidence_records_get_by_subjectreadLists all evidence records attached to a specific subject.
evidence_records_get_by_idreadRetrieves a single evidence record by its unique ID.
evidence_records_get_markdown_by_idreadRetrieves the human-readable Markdown rendering of a single evidence record. Evidence 7.289 and later.
evidence_records_prepare_statementreadPrepares a ready-to-sign in-toto statement for an artifact, build, package, release bundle or application version.
evidence_records_create_by_subjectwriteUploads an already-signed DSSE envelope as evidence and attaches it to a subject.
apptrust_create_applicationwriteCreates a new AppTrust application with its maturity level, criticality, labels and owners.
apptrust_get_application_summaryreadGets the details of an AppTrust application.
apptrust_list_applicationsreadLists AppTrust applications with filters for project, name, criticality, maturity, labels and owners.
apptrust_create_application_versionwriteCreates a new version of an application from builds, packages, artifacts, release bundles, AQL or other app versions.
apptrust_list_versionsreadLists the versions of an application, filterable by release status and tag.
apptrust_get_application_version_statusreadGets the release status for a specific application version.
apptrust_get_lifecycle_overviewreadGets the lifecycle stages configuration for a project.
apptrust_promote_versionhigh riskPromotes an application version to a lifecycle stage. Supports dry_run as a promotion type.
apptrust_get_version_promotion_historyreadGets all promotions for a specific application version.
apptrust_release_versionhigh riskReleases an application version to the PROD stage.
apptrust_rollback_versionhigh riskRolls back the latest promotion of an application version from a given stage.
apptrust_get_activity_logsreadGets activity log events from AppTrust, filterable by application, project, time range and creator.

Qué expone JFrog MCP

JFrog MCP agrupa sus 114 herramientas en 12 categorías, cada una con una política inicial sugerida.

Las 114 herramientas en JSON, listas para pegar.

Categoría	Incluye	Acceso	Política sugerida
Access	Tokens de acceso, proveedores OIDC, proyectos con sus roles, miembros y repositorios, roles globales, stages, usuarios y grupos	Lectura / Escritura	La superficie de más privilegio: crear tokens, configurar OIDC y los roles globales cambian quién puede hacer qué
Artifactory	Configuración de repositorios, versiones de paquetes y comandos de instalación, builds, salud de la federación y metadatos de almacenamiento	Lectura / Escritura	Los builds se consultan con AQL y exigen .limit(N); acota el tamaño de resultados
AppTrust	Ciclo de vida de aplicaciones, versionado, etapas, promoción, release y rollback	Lectura / Escritura	Promocionar y publicar acercan el software a producción; exige aprobación y usa dry_run antes
Evidence	Configuración del servicio, categorías de predicado y búsqueda, consulta, preparación y creación de evidencias	Lectura / Escritura	La evidencia es atestación firmada de la cadena de suministro; crearla sube un sobre DSSE
Distribution	Distribución de Release Bundle v2 a nodos edge, trackers, claves de firma y elegibilidad de edges	Lectura / Escritura	La distribución puede saltarse el bloqueo de Xray; abortar no revierte lo ya transferido
Seguridad	Resúmenes de Xray, violaciones de política, estado de indexación, búsqueda de recursos afectados y vulnerabilidades del Catalog	Lectura	Solo lectura; los hallazgos son sensibles, limita al equipo de seguridad
Curation	Comprobación de cumplimiento de paquetes de terceros, auditoría de bloqueos y de Compliant Version Selection, y solicitudes de exención	Lectura / Escritura	Decidir una exención desbloquea un paquete que la política rechazó; mantén a una persona en el bucle
Workers	Listado, creación, actualización, ejecución e historial de los JFrog Workers	Lectura / Escritura	Los Workers son TS/JS arbitrario que corre en la plataforma y puede guardar secretos; trátalo como despliegue de código
Event	Catálogo de dominios y tipos de evento de webhook, gestión de suscripciones y diagnóstico de entregas	Lectura / Escritura	Una suscripción envía eventos de la plataforma a una URL externa; force se salta la verificación
Federación de plataforma (Grid)	Topología Grid, inventario de JPD y monitorización de sincronización de entidades	Lectura	Topología y estado de sincronización de solo lectura
AI Catalog	Skills publicadas, plugins de agente y gobierno de qué servidores MCP están aprobados por proyecto	Lectura / Escritura	Aprobar un servidor MCP amplía lo que alcanza cualquier agente de ese proyecto
OneModel	Introspección del esquema y ejecución de GraphQL contra el endpoint unificado OneModel	Lectura	Introspecciona antes de ejecutar; las consultas alcanzan toda la plataforma

Conecta JFrog MCP a tu agente de IA

Cuatro pasos, menos de diez minutos.

01
Inicia sesión en NeuralTrust
Abre la plataforma de NeuralTrust, inicia sesión y ve a Getting started, después a Connect your tools.
02
Elige tu agente
Cuando te pregunte qué agente usas, elige Claude Code, Codex, Cursor, Windsurf, Copilot o tu propio código.
03
Conecta JFrog
Busca JFrog en la lista de herramientas y conéctalo.
04
Conecta tu agente
Sigue los pasos en pantalla para añadir la configuración a tu agente y envía el prompt de prueba.

Permisos hasta la herramienta individual

No permites ni bloqueas JFrog MCP entero. Permites exactamente las herramientas que puede llamar cada agente, una a una.

JFrog MCP	114 HERRAMIENTAS DISPONIBLES
Agente de revisión de código	4 herramientas, solo lectura
Bot de triaje de incidencias	9 herramientas
Asistente del equipo de plataforma	las 114 herramientas

Herramienta por herramientaConectar el servidor no es la unidad de acceso. Cada herramienta se concede por separado.
Denegado por defectoUna herramienta que no has activado nunca se expone al agente.
Aprobación en las escriturasLas herramientas que modifican o borran pueden exigir la confirmación de una persona.

¿Es seguro JFrog MCP?

Un endpoint en lugar de credenciales en claro. Cada llamada se comprueba contra la política y queda registrada.

Agentes de IASe conectan desde la configuración de tu cliente MCP.	NeuralTrustSe aplican los controles según tu configuración.	JFrog MCP114 herramientas documentadas en 23 toolsets.
Asistentes de código	Conexión y credencialespor usuario	Access23 herramientas
Copilotos internos	Políticas de acceso a herramientasdenegado por defecto	Artifactory14 herramientas
Automatizaciones de flujos	Inspección de peticiones y respuestasen línea	AppTrust12 herramientas
Clientes MCP compatibles	Monitorización de actividadalertas y registros	Resto de herramientas65 herramientas