← 返回 Skills 市场
93
总下载
0
收藏
0
当前安装
2
版本数
在 OpenClaw 中安装
/install raai-ai-support-pro
功能描述
Заменяет оператора 1-й линии на 60-80%. Полноценный AI-отдел поддержки клиентов: автоматическая категоризация обращений, SLA-трекинг с алертами, эскалационна...
安全使用建议
This package largely looks like a plausible customer-support "box" (prompts, config, docs) and could legitimately require model API keys plus optional CRM/Telegram credentials. However:
- Metadata mismatch: the registry listing said no required env vars, but SKILL.md explicitly lists ANTHROPIC_API_KEY and OPENAI_API_KEY as required and several sensitive optional creds. Ask the publisher to clarify which credentials are actually required and why two model keys are needed. Prefer one model provider or a documented fallback strategy.
- Inspect install/test scripts before running: the bundle contains install.sh and test/smoke-test.sh. Do not run them on production hosts. Review their contents line-by-line (or paste them to me for review). Run them in an isolated test environment (sandbox VM or container) with network access restricted until you confirm they behave.
- Use least-privilege/test credentials first: provide test API keys and test CRM/webhook tokens (not production keys). For Google service account JSON, create a service account with only the minimum scopes needed and revoke after testing.
- Check outbound endpoints and logging: confirm where the skill sends data (which domains/IPs). Ensure it only communicates with known model providers and your configured CRM/webhook endpoints. If install scripts upload anything to unknown endpoints, do not proceed.
- Verify provenance: there is no homepage and the source/owner are unknown. Ask the seller/owner for contact information, a code audit, or a signed release. Prefer packages with clear vendor identity or published releases.
If you want, provide the full contents of install.sh and test/smoke-test.sh and I will review them for network calls, credential exfiltration, or other red flags that would materially change this assessment.
功能分析
Type: OpenClaw Skill
Name: raai-ai-support-pro
Version: 3.5.2
The skill bundle is a comprehensive and professionally documented AI-driven customer support system. It includes detailed logic for ticket categorization, SLA tracking, and CRM integration within SKILL.md. The accompanying shell scripts (install.sh, build.sh, and smoke-test.sh) are standard utility scripts for deployment and validation, showing no signs of malicious behavior, data exfiltration, or unauthorized remote execution. All content is consistently aligned with the stated purpose of automating Tier-1 support operations.
能力标签
能力评估
Purpose & Capability
The skill claims to be a full customer-support automation with CRM and Telegram integrations; the optional env vars (TELEGRAM_BOT_TOKEN, BITRIX24_WEBHOOK_URL, AMOCRM_API_KEY, GOOGLE_SHEETS_CREDENTIALS_JSON) and the need for model API keys are coherent with that purpose. However, registry metadata earlier listed no required env vars while SKILL.md declares ANTHROPIC_API_KEY and OPENAI_API_KEY as required — this mismatch is unexplained and reduces trust in the packaging/metadata.
Instruction Scope
SKILL.md tells users to run install.sh and test/smoke-test.sh and to fill config.yaml with sensitive integration credentials; install/test scripts are present in the bundle. Because those scripts will execute on the host, they could perform arbitrary filesystem or network actions. The SKILL.md otherwise stays within support/CRM scope (categorization, SLA, templates), but the runtime instructions give a lot of implicit freedom to run scripts that are not fully shown in the provided excerpts.
Install Mechanism
There is no formal install spec in registry metadata (instruction-only), yet the package includes install.sh, build.sh and a smoke-test. The included build.sh content is benign (packaging/zip creation). Install/test scripts are present but their contents were not fully shown here — running them would write files and may make network calls. The absence of a vetted install mechanism in the registry plus executable scripts in the archive is a risk vector.
Credentials
SKILL.md requires two model API keys (ANTHROPIC_API_KEY and OPENAI_API_KEY) which can be reasonable if the product supports multiple backends, but requiring both without explanation is surprising. Optional env vars (Telegram token, CRM API keys, Google service account JSON) are proportional to CRM/notification features but are highly sensitive — the package asks users to put them into config.yaml and run install/test scripts that may access them. The registry metadata's earlier omission of these requirements is an inconsistency.
Persistence & Privilege
The skill is not marked always:true and uses default autonomous invocation settings. There's no evidence it requests persistent elevated platform privileges. However, the install/test scripts will create files and archives on disk if executed — normal for this kind of "box" product, but something to be aware of.
如何使用
- 确保已安装 OpenClaw(本地或 Docker 部署)
- 在对话框中输入安装命令:
/install raai-ai-support-pro - 安装完成后,直接呼叫该 Skill 的名称或使用
/raai-ai-support-pro触发 - 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v3.5.2
Marketplace presentation refresh.
v3.5.1
Initial marketplace release.
元数据
常见问题
Ai Support Pro 是什么?
Заменяет оператора 1-й линии на 60-80%. Полноценный AI-отдел поддержки клиентов: автоматическая категоризация обращений, SLA-трекинг с алертами, эскалационна... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 93 次。
如何安装 Ai Support Pro?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install raai-ai-support-pro」即可一键安装,无需额外配置。
Ai Support Pro 是免费的吗?
是的,Ai Support Pro 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。
Ai Support Pro 支持哪些平台?
Ai Support Pro 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。
谁开发了 Ai Support Pro?
由 RAAIPRO(@raaipro)开发并维护,当前版本 v3.5.2。
推荐 Skills