All Products
Search

This Product

    Microservices Engine:Configure RAM permissions for AI Registry

    Document Center

    Microservices Engine:Configure RAM permissions for AI Registry

    Last Updated:Jul 10, 2026

    Overview

    RAM permission policies define which AI Registry actions and resources are available to RAM users or STS temporary credentials, preventing unauthorized changes and data exposure.

    Use predefined policies for common use cases, or create custom policies for fine-grained access control.

    Prerequisites

    • An Alibaba Cloud account with AI Registry activated

    • At least one AI Registry workspace (namespace)

    • Familiarity with RAM access control, user creation, and policy configuration

    Method 1: Use predefined policies

    Attach these predefined policies directly to a RAM user or user group in the RAM console.

    Policy name

    Description

    AliyunAIRegistryReadOnlyAccess

    Grants read-only access to the console and client side. Allows viewing workspaces and prompts but not creating, modifying, or deleting resources.

    AliyunAIRegistryFullAccess

    Grants full access to all AI Registry resources.

    AliyunAIRegistryPromptFullAccess

    Grants full access to prompt resources, including creating, editing, publishing, deleting, and debugging.

    AliyunAIRegistryNamespaceFullAccess

    Grants full access to workspaces (namespaces), including creating, editing, and deleting.

    AliyunAIRegistrySkillFullAccess

    Grants full access to skill resources, including creating, editing, publishing, enabling, disabling, and deleting.

    Procedure

    1. Log on to the RAM console.

    2. In the navigation pane on the left, choose Identities > Users. Find the target RAM user.

    4. Select the required predefined policy and click OK.

    Note

    To verify, log on to the AI Registry console with the RAM user credentials and confirm that expected operations are accessible and restricted operations are blocked.

    Select AliyunAIRegistryReadOnlyAccess for view-only access, or AliyunAIRegistryFullAccess for full management.

    Method 2: Create a custom permission policy

    If predefined policies do not meet your requirements, create a custom policy for workspace-level or resource-level access control.

    2.1 Key concepts

    A custom policy consists of the following elements:

    • Effect: Allow or Deny.

    • Action: The operations to authorize.

    • Resource: The target resources.

    Action

    Actions in AI Registry are divided into two levels:

    General action—for client access:

    Action

    Description

    airegistry:Read

    Client-side read-only access for pulling resources such as prompts and skills via SDKs or APIs. Does not grant console permissions.

    airegistry:Write

    Client-side write access for resources via SDKs or APIs. Does not grant console permissions. Currently, only skills support client-side writes.

    API-level actions—for console operations (fine-grained authorization by specific API):

    Category

    Action

    Description

    Workspace

    airegistry:CreateNamespace

    Creates a workspace.

    airegistry:ListNamespaces

    Lists workspaces.

    airegistry:GetNamespace

    Gets workspace details.

    airegistry:UpdateNamespace

    Updates a workspace.

    airegistry:DeleteNamespace

    Deletes a workspace.

    airegistry:CreateNamespaceWithSource

    Creates a workspace with a specified source.

    airegistry:ListNamespaceBySource

    Lists workspaces filtered by source.

    Prompt

    airegistry:ListPrompts

    Lists prompts.

    airegistry:GetPrompt

    Gets prompt details.

    airegistry:CreatePrompt

    Creates a prompt.

    airegistry:UpdatePrompt

    Updates a prompt.

    airegistry:DeletePrompt

    Deletes a prompt.

    airegistry:ListPromptVersions

    Lists prompt versions.

    airegistry:GetPromptVersion

    Gets prompt version information.

    airegistry:GetPromptVersionDetail

    Gets prompt version details.

    airegistry:CreatePromptVersion

    Creates a prompt version.

    airegistry:UpdatePromptVersion

    Updates a prompt version.

    airegistry:SubmitPromptVersion

    Publishes a prompt version.

    airegistry:DeletePromptVersion

    Deletes a prompt version.

    airegistry:GetPromptGovernance

    Gets prompt governance information.

    Debugging and optimization

    airegistry:InvokePromptDebugStream

    Debugs a prompt (streaming).

    airegistry:InvokePromptOptimizeStream

    Optimizes a prompt (streaming).

    airegistry:GetSupportedModels

    Gets the list of supported models.

    Skill

    airegistry:CreateSkillDraft

    Creates a skill draft.

    airegistry:UpdateSkillDraft

    Updates a skill draft.

    airegistry:DeleteSkillDraft

    Deletes a skill draft.

    airegistry:SubmitSkillVersion

    Submits a skill version.

    airegistry:PublishSkillVersion

    Publishes a skill version.

    airegistry:ForcePublishSkillVersion

    Force-publishes a skill version.

    airegistry:UpdateSkillBizTags

    Updates the business tags of a skill.

    airegistry:UpdateSkillLabels

    Updates the labels of a skill.

    airegistry:OnlineSkill

    Enables a skill.

    airegistry:OfflineSkill

    Disables a skill.

    airegistry:UpdateSkillScope

    Updates the scope of a skill.

    airegistry:GetSkillDetail

    Gets skill details.

    airegistry:GetSkillVersionDetail

    Gets skill version details.

    airegistry:DownloadSkillVersion

    Downloads a skill version.

    airegistry:ListSkills

    Lists skills.

    airegistry:DeleteSkill

    Deletes a skill.

    airegistry:UploadSkill

    Uploads a skill.

    airegistry:UploadSkillViaOss

    Uploads a skill via OSS.

    airegistry:DownloadSkillVersionViaOss

    Downloads a skill version via OSS.

    Note

    • airegistry:Read and airegistry:Write grant client-side access only. Neither grants console permissions. To allow both client and console access, combine airegistry:Read and airegistry:Write with API-level actions.

    • Alternatively, use airegistry:* as the Action to grant both client-side read/write access and all API-level operation permissions.

    Resource

    Resources are identified by Alibaba Cloud Resource Name (ARN):

    acs:airegistry:*:*:instance/saas/{workspace-id}/{domain}/{resource-type}/{resource-name}

    Where:

    • workspace-id: Your workspace ID from the AI Registry console.

    • domain: Currently fixed to DEFAULT.

    • resource-type: The resource type, such as prompt, skill, mcp, or agent.

    • resource-name: The name of the specific resource.

    Use the wildcard (*) to match multiple resources.

    2.2 Authorization granularity

    Control access at different levels of granularity:

    Scope of authorization

    Example resource

    All resources in all workspaces

    acs:airegistry:*:*:instance/saas/*

    All resources in a specific workspace

    acs:airegistry:*:*:instance/saas/{your-workspace-id}/*

    A specific type of resource in a specific workspace

    acs:airegistry:*:*:instance/saas/{your-workspace-id}/DEFAULT/prompt/*

    A specific type of resource in a specific workspace (skill)

    acs:airegistry:*:*:instance/saas/{your-workspace-id}/DEFAULT/skill/*

    A specific resource

    acs:airegistry:*:*:instance/saas/{your-workspace-id}/DEFAULT/prompt/my-prompt

    2.3 Create a custom policy

    1. Log on to the RAM console.

    2. In the navigation pane on the left, choose Permissions > Policies.

    3. Click Create Policyand select the JSON Editor tab.

    4. Write the policy content by referring to the examples below and click OK.

    5. Attach the created policy to the target RAM user or user group.

    Note

    To verify, log on to the AI Registry console with the RAM user credentials and confirm that expected operations are accessible and restricted operations are blocked.

    2.4 Common scenarios

    Scenario 1: Grant read-only client access

    An application pulls resources such as prompts via an SDK or API and requires only read access.

    {
  "Statement": [
    {
      "Effect": "Allow",
      "Action": "airegistry:Read",
      "Resource": "acs:airegistry:*:*:instance/saas/{your-workspace-id}/*"
    }
  ],
  "Version": "1"
}

    Scenario 2: Grant full management of a workspace

    A team member needs full management permissions (console and client) for a specific workspace.

    {
  "Statement": [
    {
      "Effect": "Allow",
      "Action": "airegistry:*",
      "Resource": "acs:airegistry:*:*:instance/saas/{your-workspace-id}/*"
    }
  ],
  "Version": "1"
}

    Scenario 3: Restrict access to prompts

    A prompt engineer needs to manage prompts but should not access other resource types.

    {
  "Statement": [
    {
      "Effect": "Allow",
      "Action": [
        "airegistry:Read",
        "airegistry:GetPrompt",
        "airegistry:CreatePrompt",
        "airegistry:UpdatePrompt",
        "airegistry:DeletePrompt",
        "airegistry:ListPromptVersions",
        "airegistry:GetPromptVersion",
        "airegistry:GetPromptVersionDetail",
        "airegistry:CreatePromptVersion",
        "airegistry:UpdatePromptVersion",
        "airegistry:SubmitPromptVersion",
        "airegistry:DeletePromptVersion",
        "airegistry:GetPromptGovernance",
        "airegistry:InvokePromptDebugStream",
        "airegistry:InvokePromptOptimizeStream"
      ],
      "Resource": "acs:airegistry:*:*:instance/saas/{your-workspace-id}/DEFAULT/prompt/*"
    },
    {
      "Effect": "Allow",
      "Action": [
        "airegistry:ListNamespaces",
        "airegistry:ListNamespaceBySource",
        "airegistry:GetNamespace",
        "airegistry:GetSupportedModels"
      ],
      "Resource": "acs:airegistry:*:*:instance/saas/*"
    },
    {
      "Effect": "Allow",
      "Action": "airegistry:ListPrompts",
      "Resource": "acs:airegistry:*:*:instance/saas/{your-workspace-id}/DEFAULT/prompt"
    }
  ],
  "Version": "1"
}
    Note

    ListPrompts is a collection-level operation, so its Resource is .../prompt (without /*).

    Alternatively, use the predefined policy AliyunAIRegistryPromptFullAccess to achieve the same result.

    Scenario 4: Combine client and console permissions

    An application pulls resources via a client, and the user also manages prompts in the console.

    {
  "Statement": [
    {
      "Effect": "Allow",
      "Action": "airegistry:Read",
      "Resource": "acs:airegistry:*:*:instance/saas/{your-workspace-id}/*"
    },
    {
      "Effect": "Allow",
      "Action": [
        "airegistry:CreatePrompt",
        "airegistry:UpdatePrompt",
        "airegistry:DeletePrompt",
        "airegistry:CreatePromptVersion",
        "airegistry:UpdatePromptVersion",
        "airegistry:SubmitPromptVersion",
        "airegistry:DeletePromptVersion",
        "airegistry:InvokePromptDebugStream",
        "airegistry:InvokePromptOptimizeStream"
      ],
      "Resource": "acs:airegistry:*:*:instance/saas/{your-workspace-id}/DEFAULT/prompt/*"
    }
  ],
  "Version": "1"
}

    Scenario 5: Restrict access to skills

    A team member needs to manage skills but should not access other resource types.

    {
  "Statement": [
    {
      "Effect": "Allow",
      "Action": [
        "airegistry:Read",
        "airegistry:Write",
        "airegistry:UpdateSkillDraft",
        "airegistry:DeleteSkillDraft",
        "airegistry:SubmitSkillVersion",
        "airegistry:PublishSkillVersion",
        "airegistry:ForcePublishSkillVersion",
        "airegistry:UpdateSkillBizTags",
        "airegistry:UpdateSkillLabels",
        "airegistry:OnlineSkill",
        "airegistry:OfflineSkill",
        "airegistry:UpdateSkillScope",
        "airegistry:GetSkillDetail",
        "airegistry:GetSkillVersionDetail",
        "airegistry:DownloadSkillVersion",
        "airegistry:DeleteSkill",
        "airegistry:DownloadSkillVersionViaOss"
      ],
      "Resource": "acs:airegistry:*:*:instance/saas/{your-workspace-id}/DEFAULT/skill/*"
    },
    {
      "Effect": "Allow",
      "Action": [
        "airegistry:ListNamespaces",
        "airegistry:ListNamespaceBySource",
        "airegistry:GetNamespace"
      ],
      "Resource": "acs:airegistry:*:*:instance/saas/*"
    },
    {
      "Effect": "Allow",
      "Action": [
        "airegistry:ListSkills",
        "airegistry:CreateSkillDraft",
        "airegistry:UploadSkill",
        "airegistry:UploadSkillViaOss"
      ],
      "Resource": "acs:airegistry:*:*:instance/saas/{your-workspace-id}/DEFAULT/skill"
    }
  ],
  "Version": "1"
}
    Note

    ListSkills, CreateSkillDraft, UploadSkill, and UploadSkillViaOss are collection-level operations, so their Resource is .../skill (without /*).

    Alternatively, use the predefined policy AliyunAIRegistrySkillFullAccess to achieve the same result.

    Scenario 6: Scope permissions to a specific resource

    Use case: Allow a RAM user to access only one specific prompt.

    {
  "Statement": [
    {
      "Effect": "Allow",
      "Action": "airegistry:*",
      "Resource": "acs:airegistry:*:*:instance/saas/{your-workspace-id}/DEFAULT/prompt/my-prompt"
    }
  ],
  "Version": "1"
}

    2.5 Usage notes

    • Deny overrides Allow: When both Allow and Deny apply to the same action, Deny takes precedence. Use this to grant broad permissions while restricting specific actions.

    • Workspace isolation: AI Registry enforces workspace ownership. You cannot access workspaces owned by other Alibaba Cloud accounts, even if specified in a RAM policy.

    • Workspace ID in policies: Replace {your-workspace-id} with your actual workspace ID from the AI Registry console.

    • Principle of least privilege: Grant each user only the minimum permissions required for their tasks. Avoid overly broad wildcards.

    FAQ

    Q: What should I do if a RAM user receives a "permission denied" error?

    A: Check the following:

    • Verify that the correct permission policy is attached to the RAM user.

    • Verify that the workspace ID in the permission policy matches the workspace being accessed.

    • Verify that the Action covers the required operation type (airegistry:Read, airegistry:Write, or specific API-level actions).

    • If you are using a Deny policy, check that it is not unintentionally blocking required operations.

    Q: How do I allow a RAM user to access only specific workspaces?

    A: In the Resource element, specify exact workspace IDs instead of the wildcard (*). For multiple workspaces, add multiple Statement elements.

    Q: How do I configure permissions for access using STS temporary credentials?

    A: STS temporary credentials inherit permissions from two sources: the policy attached to the RAM role, and the policy specified in the AssumeRole API call. Ensure the RAM role has the required AI Registry permissions and the AssumeRole policy does not restrict them.