Zalo

OpenClaw skill for Zalo Bot API workflows (bot token) plus optional guidance on unofficial personal automation tools.

This skill appears to be a legitimate Zalo bot guidance bundle, but it has two practical issues to consider before installing: (1) the skill's documents expect you to supply sensitive secrets (e.g., ZALO_BOT_TOKEN, optional tokenFile, webhookSecret), yet the registry metadata does not declare these required environment variables — verify what secrets the skill will actually read and how they will be provided; (2) the included 'personal automation' notes reference cookies and device identifiers (sensitive session state) and are explicitly unofficial — avoid using that branch in production and do not store cookies on shared hosts. Because this is instruction-only, there's no code to audit here, so confirm with the publisher (or view an implementation) how tokens and tokenFiles are read, whether anything will be transmitted to third-party endpoints, and whether the agent will be given those secrets for autonomous invocation. If the publisher updates the metadata to declare required env vars and provides an implementation you can inspect, re-evaluate; until then treat token/cookie provisioning conservatively and limit scope (use dev tokens, allowlists, rotate tokens).

Type: OpenClaw Skill Name: zalo Version: 1.0.1 The skill is classified as suspicious due to its detailed guidance on 'unofficial personal automation' methods that involve handling highly sensitive 'cookies and device identifiers' for Zalo Web, as outlined in `references/zalo-personal-zca-js.md`. While the documentation explicitly warns about the risks (violating platform terms, account bans, sensitivity of cookies), providing instructions on how to manage such sensitive, non-API-based authentication data for potentially unauthorized automation introduces a significant security risk. The mention of `zca-js` and `zca-cli` also points to reliance on external, unofficial tools.