alibabacloud-dms-data-agent-platform-setup

Create Dify Instance

Provision a Dify instance automatically via Alibaba Cloud OpenAPI. Supports Simple Mode (create all resources from scratch) and Advanced Mode (fine-grained control over each component).

Prerequisites

1. Check Aliyun CLI

[REQUIRED] Verify Aliyun CLI version >= 3.3.1 before proceeding.

aliyun version

If the command is not found or the version is below 3.3.1, install or upgrade:

macOS (Homebrew, recommended)

brew install aliyun-cli
# Upgrade if already installed
brew upgrade aliyun-cli

After installation, enable automatic plugin installation:

aliyun configure set --auto-plugin-install true

2. Observability — Session ID & User-Agent

[REQUIRED] Generate a session-id at the start of every skill session. Use it consistently across ALL CLI commands and Python SDK calls.

Session-ID Generation Rule

• Generate once per session, reuse throughout the entire workflow
• Format: 32-character lowercase hex string (e.g., openssl rand -hex 16)
• Must be consistent across CLI / SDK / Terraform within the same session

export SKILL_SESSION_ID=$(openssl rand -hex 16)

User-Agent Template

AlibabaCloud-Agent-Skills/alibabacloud-dms-data-agent-platform-setup/{session-id}

• Python scripts (detect_dify_instance.py, openAPI_call.py): read SKILL_SESSION_ID from environment automatically
• Aliyun CLI commands: inject via --user-agent flag on each command:

aliyun dms-enterprise <action> \
  --endpoint dms-enterprise.aliyuncs.com \
  --user-agent "AlibabaCloud-Agent-Skills/alibabacloud-dms-data-agent-platform-setup/${SKILL_SESSION_ID}"

Do NOT use aliyun configure ai-mode set-user-agent (deprecated). Always use per-command --user-agent injection.

3. Configure Alibaba Cloud Credentials

[REQUIRED] Both credential sets must be configured — Aliyun CLI credentials AND Python SDK credentials. Neither can be omitted. NEVER read, echo, or print AK/SK values.

This skill performs two types of operations, each using a different credential method:

• Query instance list (aliyun dms-enterprise list-instances): uses Aliyun CLI credentials
• Provision Dify instance (openAPI_call.py): uses the Alibaba Cloud default credential chain

3a. Configure Aliyun CLI Credentials

aliyun configure list

Confirm that a valid profile exists in the output (AK, STS, or OAuth).

If no valid profile exists, stop and prompt the user to:

1. Obtain an AccessKey from the Alibaba Cloud Console
2. Configure credentials outside of this session to avoid exposing secrets:

   aliyun configure set \
     --mode AK \
     --access-key-id <your-access-key-id> \
     --access-key-secret <your-access-key-secret> \
     --region cn-hangzhou
   

3. Re-run aliyun configure list to confirm the profile is active

3b. Configure Python Script Credentials

openAPI_call.py uses the Alibaba Cloud default credential chain — no environment variables need to be set manually. The SDK automatically resolves credentials in the following order: environment variables, credentials file, instance RAM role, etc.

Configure your credentials by following the official guide: Alibaba Cloud Python SDK v2 — Manage Access Credentials

NEVER hardcode AK/SK values in code or pass them as command-line arguments.

4. Python Environment

It is recommended to use uv to create an isolated virtual environment with pinned dependencies:

uv venv .venv
uv pip install --python .venv/bin/python -r scripts/requirements.txt

requirements.txt is provided in ./scripts/.

Script Location

• ./scripts/detect_dify_instance.py — Check whether the DMS Dify service is available in the target region
• ./scripts/openAPI_call.py — Provision a Dify instance

Run commands from the directory containing this SKILL.md file.

---

Pre-check: Verify DMS Service Availability

[REQUIRED] Before running openAPI_call.py, run detect_dify_instance.py to verify that the DMS service is accessible in the target region.

.venv/bin/python ./scripts/detect_dify_instance.py <region-id>

Example:

.venv/bin/python ./scripts/detect_dify_instance.py cn-shanghai

• If the API returns an empty list ([]), that is normal — it means no Dify instances exist yet, but the service is available.
• If the API returns an error, prompt the user to enable the DMS service in the Alibaba Cloud Console before proceeding.

---

Simple Mode

All components (Workspace, database, KV store, vector database) are newly created.

Parameters to Collect from User

Execution Command

.venv/bin/python ./scripts/openAPI_call.py '{
    "VpcId": "<VpcId>",
    "VSwitchId": "<VSwitchId>",
    "BackupVSwitchId": "<BackupVSwitchId>",
    "SecurityGroupId": "<SecurityGroupId>",
    "ZoneId": "<ZoneId>",
    "DataRegion": "<DataRegion>",
    "ResourceQuota": "12CU",
    "WorkspaceOption": "CreateNewInstance",
    "WorkspaceName": "<WorkspaceName>",
    "DatabaseOption": "CreateNewInstance",
    "DbInstanceAccount": "<account>",
    "DbInstancePassword": "<password>",
    "KvStoreOption": "CreateNewInstance",
    "KvStoreAccount": "<account>",
    "KvStorePassword": "<password>",
    "VectordbOption": "CreateNewInstance",
    "VectordbAccount": "<account>",
    "VectordbPassword": "<password>",
    "StorageType": "cloud_essd",
    "NatGatewayOption": "NoNeed",
    "MajorVersion": "1.13.x",
    "Edition": "OpenCommunity",
    "DryRun": true
}'

---

Advanced Mode

Allows fine-grained control over all parameters, including using existing Workspace, database, KV store, and vector database instances.

Step 1: Collect Base Network Parameters

Ask the user for the following:

Step 2: Confirm WorkspaceOption

Ask the user: use an existing Workspace or create a new one?

• UseExistingInstance: user must provide WorkspaceId (string)
• CreateNewInstance: user must provide WorkspaceName
• Note: WorkspaceId and WorkspaceName are mutually exclusive. If both are provided, prompt the user to correct the input.

Step 3: Confirm Each Sub-Service Option Individually

Ask the user to choose for each of the following independently:

DatabaseOption

• CreateNewInstance: no additional parameters needed; uses default configuration
• UseExistingInstance:
  • Ask the user if they know the DbResourceId (integer)
  • If not, run the following command and find the InstanceId from InstanceList.Instance:

    aliyun dms-enterprise list-instances \
      --endpoint dms-enterprise.aliyuncs.com \
      --user-agent "AlibabaCloud-Agent-Skills/alibabacloud-dms-data-agent-platform-setup/${SKILL_SESSION_ID}"
    

KvStoreOption

• CreateNewInstance: no additional parameters needed
• UseExistingInstance:
  • Ask the user if they know the KvStoreResourceId (integer)
  • If not, run:

    aliyun dms-enterprise list-instances \
      --endpoint dms-enterprise.aliyuncs.com \
      --user-agent "AlibabaCloud-Agent-Skills/alibabacloud-dms-data-agent-platform-setup/${SKILL_SESSION_ID}"
    

VectordbOption

• CreateNewInstance: no additional parameters needed
• UseExistingInstance:
  • Ask the user if they know the VectordbResourceId (integer)
  • If not, run:

    aliyun dms-enterprise list-instances \
      --endpoint dms-enterprise.aliyuncs.com \
      --user-agent "AlibabaCloud-Agent-Skills/alibabacloud-dms-data-agent-platform-setup/${SKILL_SESSION_ID}"
    

Step 4: Collect Account and Password

Ask the user for account name and password, and fill in:

• DbInstanceAccount / DbInstancePassword
• KvStoreAccount / KvStorePassword
• VectordbAccount / VectordbPassword

Step 5: Confirm Other Advanced Parameters

The following parameters have default values. Ask the user if any need to be changed:

Step 6: Dry Run and Final Provisioning

1. Construct the full JSON with DryRun=true and run the script
2. After confirming no errors, set DryRun to false and run again to provision

---

Notes

• The Python script uses the Alibaba Cloud default credential chain; configure credentials per the official guide before running the script
• WorkspaceId (use existing Workspace) and WorkspaceName (create new Workspace) are mutually exclusive
• DbResourceId, KvStoreResourceId, and VectordbResourceId are all integer types
• Always perform a dry run with DryRun=true before final provisioning
• Ensure SKILL_SESSION_ID is exported before running any command in this workflow