← Back to Skills Marketplace
sdk-team

Alibabacloud Tablestore Openclaw Memory

by alibabacloud-skills-team · GitHub ↗ · v0.0.1 · MIT-0
cross-platform ⚠ suspicious
76
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install alibabacloud-tablestore-openclaw-memory
Description
This skill installs and configures the **Tablestore Mem0** plugin for OpenClaw. Tablestore Mem0 uses Alibaba Cloud Tablestore as the vector store backend for...
README (SKILL.md)

Prerequisites

The user must have:

  1. An Alibaba Cloud account
  2. CredentialsAccessKey ID + Secret (with AliyunOTSFullAccess permission) or ECS RAM Role (recommended on ECS)
  3. An 阿里云百炼 API Key (for Qwen LLM and embedding)
  4. Optional: existing Tablestore instance endpoint + instance name (otherwise auto-created)
  5. Optional: region ID for auto-provisioning (defaults to cn-hangzhou)

📋 RAM Permissions: See references/ram-policies.md for the complete list of required API permissions.


Definition of Done

This task is NOT complete until all of the following are true:

  1. User has provided credentials
  2. Plugin is installed (version 0.8.2)
  3. openclaw.json is configured correctly
  4. OpenClaw is restarted
  5. Setup is verified (plugin loads without errors)

Onboarding

Step 0 — Collect credentials

[AGENT] Ask the user:

To set up Tablestore Mem0, I need:

Required:

  1. Alibaba Cloud AccessKey ID
  2. Alibaba Cloud AccessKey Secret
  3. 阿里云百炼 API Key

Optional (if you have an existing Tablestore instance): 4. Tablestore endpoint (e.g. https://my-instance.cn-hangzhou.ots.aliyuncs.com) 5. Tablestore instance name

If you don't provide an endpoint and instance name, I'll automatically create a new Tablestore instance. In that case, I need: 6. Region ID (e.g. cn-hangzhou, cn-shanghai)

Ask the user:

Question 1 — Authentication: How would you like to authenticate?

  • Option A: ECS RAM Role (recommended on ECS) — Provide the role name.
  • Option B: AccessKey — Provide AccessKey ID and Secret.

Question 2 — Tablestore instance: Do you already have a Tablestore instance?

  • Yes → Provide endpoint and instance name.
  • No → Provide region ID for auto-provisioning.

⚠️ HIGH-RISK OPERATION — Pre-flight Confirmation:

If the user answers No, the agent MUST:

  1. Explicit Confirmation:

    ⚠️ This setup will create a new Tablestore instance in your Alibaba Cloud account.

    • Instance type: VCU (pay-as-you-go, typically under ¥1/month)
    • Region: as specified Do you confirm? (yes/no)
  2. Double-check:

    Please verify in Tablestore Console (https://ots.console.aliyun.com/) that you don't have existing instances.

Do NOT proceed until the user explicitly confirms.

If the user is missing credentials, point them to:

  • Alibaba Cloud sign-up: https://account.alibabacloud.com/register/intl_register.htm
  • AccessKey creation: https://ram.console.aliyun.com/manage/ak
  • 阿里云百炼 API Key: https://dashscope.console.aliyun.com/apiKey

Setting up ECS RAM Role (if the user chose Option A)

  1. Attach a RAM role to the ECS instanceDocumentation
  2. Grant the role AliyunOTSFullAccessReference

Step 1 — Install plugin

Plugin version: 0.8.2

⚠️ External Package Installation Notice:

This skill will install @tablestore/[email protected] from npm registry. By proceeding, you acknowledge installing an external package. Do you consent? (yes/no)

Wait for user confirmation, then:

NPMJS_TIME="$(curl -o /dev/null -sS --connect-timeout 2 --max-time 6 -w '%{time_total}' https://registry.npmjs.org/@tablestore%2fopenclaw-mem0 || echo timeout)"
MIRROR_TIME="$(curl -o /dev/null -sS --connect-timeout 2 --max-time 6 -w '%{time_total}' https://registry.npmmirror.com/@tablestore%2fopenclaw-mem0 || echo timeout)"

if [ "$MIRROR_TIME" != "timeout" ] && { [ "$NPMJS_TIME" = "timeout" ] || awk "BEGIN { exit !($NPMJS_TIME > 2 && $MIRROR_TIME \x3C $NPMJS_TIME) }"; }; then
  echo "Using China npm mirror"
  NPM_CONFIG_REGISTRY=https://registry.npmmirror.com openclaw plugins install @tablestore/[email protected]
else
  openclaw plugins install @tablestore/[email protected]
fi

If extract tar timed out, fallback to manual install:

cd /tmp && npm pack @tablestore/[email protected]
mkdir -p /tmp/openclaw-mem0-install
tar xzf /tmp/tablestore-openclaw-mem0-*.tgz -C /tmp/openclaw-mem0-install --strip-components=1
openclaw plugins install /tmp/openclaw-mem0-install

Step 2 — Detect OpenClaw version

openclaw --version
  • Version >= 2.2.0 → use Step 3A
  • Version \x3C 2.2.0 → use Step 3B (remove .plugins.allow line)

Step 3 — Configure openclaw.json

Detect config file path:

OPENCLAW_CONFIG="$(openclaw config file 2>/dev/null | head -1)"
OPENCLAW_CONFIG="${OPENCLAW_CONFIG/#\~/$HOME}"
echo "Config file: $OPENCLAW_CONFIG"

⚠️ Credential Security (REQUIRED):

ALWAYS prefer environment variables:

export TABLESTORE_ACCESS_KEY_ID="\x3Cyour-access-key-id>"
export TABLESTORE_ACCESS_KEY_SECRET="\x3Cyour-access-key-secret>"
export DASHSCOPE_API_KEY="\x3Cyour-dashscope-api-key>"

⚠️ Input Validation (REQUIRED):

Input Validation Rule
AccessKey ID Regex: ^LTAI[Nt][A-Za-z0-9]{12,20}$
AccessKey Secret Alphanumeric, ~30 chars
Region ID Regex: ^cn-[a-z]+(-[0-9]+)?$ or ^[a-z]+-[a-z]+-[0-9]+$
Endpoint Regex: ^https://[a-z0-9-]+\.[a-z0-9-]+\.ots\.aliyuncs\.com$
Instance Name Regex: ^[a-z][a-z0-9-]{0,19}$

⚠️ Embedder/LLM provider MUST be "openai" with baseURL "https://dashscope.aliyuncs.com/compatible-mode/v1", NOT "dashscope".

Step 3A — OpenClaw ≥2.2.0

Auth method Has existing instance? vectorStore.config fields
AccessKey No accessKeyId, accessKeySecret, regionId
AccessKey Yes accessKeyId, accessKeySecret, endpoint, instanceName
ECS RAM Role No roleName, regionId
ECS RAM Role Yes roleName, endpoint, instanceName

Template (AccessKey auth + auto-create):

jq '
  .plugins.slots.memory = "openclaw-mem0" |
  .plugins.entries["openclaw-mem0"] = {
    enabled: true,
    config: {
      mode: "open-source",
      oss: {
        vectorStore: {
          provider: "tablestore",
          config: {
            accessKeyId: "${TABLESTORE_ACCESS_KEY_ID}",
            accessKeySecret: "${TABLESTORE_ACCESS_KEY_SECRET}",
            regionId: "cn-hangzhou"
          }
        },
        embedder: {
          provider: "openai",
          config: {
            apiKey: "${DASHSCOPE_API_KEY}",
            model: "text-embedding-v3",
            baseURL: "https://dashscope.aliyuncs.com/compatible-mode/v1"
          }
        },
        llm: {
          provider: "openai",
          config: {
            apiKey: "${DASHSCOPE_API_KEY}",
            model: "qwen-plus",
            baseURL: "https://dashscope.aliyuncs.com/compatible-mode/v1"
          }
        }
      }
    }
  } |
  .plugins.allow = ((.plugins.allow // []) + ["openclaw-mem0"] | unique)
' "$OPENCLAW_CONFIG" > /tmp/openclaw-tmp.json && mv /tmp/openclaw-tmp.json "$OPENCLAW_CONFIG"

For existing instance: replace regionId with endpoint and instanceName. For ECS RAM Role: replace accessKeyId/accessKeySecret with roleName.

Step 3B — OpenClaw \x3C2.2.0

Same as Step 3A but remove the .plugins.allow line.

Step 4 — Restart OpenClaw

openclaw gateway restart

Wait ~1 minute for the gateway to restart.

Step 5 — Verify setup

openclaw mem0 stats

Success criteria: plugin loads without errors, shows mode and memory count.

Step 6 — Handoff

✅ Tablestore Mem0 is ready.

Your memory is now backed by Alibaba Cloud Tablestore.
The plugin will automatically recall relevant memories before each conversation
and capture key facts after each conversation.

Manual commands: "remember this" / "what do you know about me?" / "forget that"

Configuration: Tablestore vector store + 阿里云百炼 text-embedding-v3 + Qwen-Plus
Instance: \x3Cauto-created or user-provided>

Security: Prefer ECS RAM Role on ECS. Never commit credentials to version control.

Configuration Reference

Parameter Required Default Description
accessKeyId No* AccessKey ID (or env TABLESTORE_ACCESS_KEY_ID)
accessKeySecret No* AccessKey Secret (or env TABLESTORE_ACCESS_KEY_SECRET)
roleName No* ECS RAM Role name (or env TABLESTORE_ROLE_NAME)
endpoint No Auto-created Tablestore instance endpoint
instanceName No Auto-created Tablestore instance name
regionId No cn-hangzhou Region for auto-provisioning
tableName No mem0_vectors Data table name
dimension No 1024 Vector dimension

*Required unless roleName is set.

Auto-provisioning

When endpoint/instanceName not provided, creates VCU instance with:

  • VCU=0 (pay-as-you-go)
  • Auto-enables public internet access if VPC unreachable

Troubleshooting

Symptom Fix
Plugin not loading Check slots.memory = "openclaw-mem0" and enabled = true
AccessKeyId is invalid Verify AccessKey is correct and account is active
Auto-provisioning fails Ensure AccessKey has AliyunOTSFullAccess
Embedding errors Verify DASHSCOPE_API_KEY is set
VPC endpoint unreachable Plugin auto-enables public access; verify AliyunOTSFullAccess
extract tar timed out Use manual fallback: npm pack + system tar
Provider shows dashscope Change provider to "openai" with correct baseURL

Update

Do not set up automatic updates. Only update when explicitly requested.

Usage Guidance
This skill appears to do what it says, but take these precautions before installing: (1) Confirm you want the agent to create cloud resources—do not proceed until you explicitly allow auto-provisioning. (2) Prefer using an ECS RAM role with least-privilege policies over long-lived AccessKeys; if you must use AccessKeys, create a limited service account and avoid giving full account-wide privileges. (3) Review the npm package @tablestore/[email protected] (author, homepage, and recent releases) before installing; installing unreviewed packages at runtime is a supply-chain risk. (4) Be aware the setup may enable public internet access for a Tablestore instance if VPC endpoints fail—this increases exposure. (5) Never paste secrets into chat; use the agent's interactive prompt to supply credentials and prefer environment variables or short-lived credentials. (6) Because the skill metadata lacks a homepage and the publisher is unknown, exercise additional caution and verify package provenance and the RAM policy requirements before proceeding.
Capability Analysis
Type: OpenClaw Skill Name: alibabacloud-tablestore-openclaw-memory Version: 0.0.1 The skill bundle automates the installation of an external npm package (@tablestore/openclaw-mem0) and requires the user to provide high-privilege Alibaba Cloud AccessKeys and DashScope API keys. It is classified as suspicious because it handles these sensitive credentials and provides instructions for the agent to write them into a local configuration file (openclaw.json) using jq (SKILL.md). While the skill includes confirmation prompts for high-risk operations and input validation, the process of embedding raw secrets into configuration files and executing shell-based installation logic represents a significant security risk, even if aligned with the stated purpose of setting up cloud-backed memory.
Capability Tags
cryptocan-make-purchases
Capability Assessment
Purpose & Capability
Name/description (Tablestore mem0 plugin for OpenClaw) match the actions the SKILL.md describes: installing a plugin, configuring openclaw.json, and creating/using an Alibaba Cloud Tablestore instance. Required inputs (AccessKey or ECS RAM role and Dashscope API key) are appropriate for provisioning Tablestore and obtaining embeddings from阿里云百炼/Qwen.
Instruction Scope
Instructions remain within the plugin setup scope (collect credentials interactively, optionally create a Tablestore instance, install plugin, edit openclaw.json, restart OpenClaw, verify). The agent is explicit about requiring user confirmation before auto-provisioning. It will modify OpenClaw config and restart the service—expected for this task—but will also potentially enable public internet access on instances (via UpdateInstance) if VPC endpoints are unreachable, which has security implications.
Install Mechanism
This is instruction-only but tells the agent to install an external npm package (@tablestore/[email protected]) from the public registry (with fallback to tar extraction). Installing third-party packages at runtime is a standard approach for plugins but carries the usual supply-chain risk; the SKILL.md does not include a vetted source or homepage to audit the package contents.
Credentials
Requested credentials (Alibaba AccessKey ID/Secret or ECS RAM role, plus Dashscope API key) are relevant to the stated purpose. However, the SKILL requires broad OTS permissions (documented as AliyunOTSFullAccess / CreateInstance, UpdateInstance, etc.), which are high privilege. The skill recommends environment variables for secrets (good), but granting full Tablestore permissions or enabling public access should be considered carefully.
Persistence & Privilege
Skill is not marked always:true and is user-invocable; it will modify its own runtime environment (openclaw.json) and restart OpenClaw as part of setup, which is consistent with a plugin installer. No evidence it attempts to persist beyond normal plugin configuration or alter other skills' configs.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install alibabacloud-tablestore-openclaw-memory
  3. After installation, invoke the skill by name or use /alibabacloud-tablestore-openclaw-memory
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.0.1
Initial release of Alibaba Cloud Tablestore Mem0 integration for OpenClaw. - Enables AI agents in OpenClaw to store and retrieve long-term memory using Alibaba Cloud Tablestore as a persistent vector backend. - Guides users step-by-step through credential collection, plugin installation (v0.8.2), and configuration, with strong input validation and risk confirmation for resource provisioning. - Supports both AccessKey and ECS RAM Role authentication, and either pre-existing or auto-created Tablestore instances. - Ensures secure credential use and provides onboarding prompts for missing prerequisites. - Includes explicit user confirmations before installing external packages or provisioning cloud resources. - Detailed instructions for openclaw.json configuration and plugin verification for OpenClaw versions ≥2.2.0 and <2.2.0.
Metadata
Slug alibabacloud-tablestore-openclaw-memory
Version 0.0.1
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is Alibabacloud Tablestore Openclaw Memory?

This skill installs and configures the **Tablestore Mem0** plugin for OpenClaw. Tablestore Mem0 uses Alibaba Cloud Tablestore as the vector store backend for... It is an AI Agent Skill for Claude Code / OpenClaw, with 76 downloads so far.

How do I install Alibabacloud Tablestore Openclaw Memory?

Run "/install alibabacloud-tablestore-openclaw-memory" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Alibabacloud Tablestore Openclaw Memory free?

Yes, Alibabacloud Tablestore Openclaw Memory is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Alibabacloud Tablestore Openclaw Memory support?

Alibabacloud Tablestore Openclaw Memory is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Alibabacloud Tablestore Openclaw Memory?

It is built and maintained by alibabacloud-skills-team (@sdk-team); the current version is v0.0.1.

💬 Comments