Install
openclaw skills install @huaweiclouddev/huawei-cloud-icp-rule-consultICP Filing (备案) rule consultation skill for Huawei Cloud. Based on Python 3.9+ with requests/Playwright, answers whether ICP filing is required, filing process & materials, account & entity limits, change filing, access filing, cancellation filing, APP filing, and migration/transfer filing. Read-only consultation, no filing operations on behalf of users. Use only when the user explicitly mentions 备案/ICP备案/备案规则/备案流程/接入备案/变更备案/注销备案/APP备案/迁移备案; refuses domain registration, DNS configuration, server purchase, real-name authentication, pricing. Trigger: 备案, ICP备案, 备案规则, 备案流程, 接入备案, 变更备案, 注销备案, APP备案, 迁移备案, 转移备案, 管局要求, 备案材料, 备案授权码
openclaw skills install @huaweiclouddev/huawei-cloud-icp-rule-consultHuawei Cloud ICP Filing — Read-Only Rule Consultation
Using embedded knowledge base (130+ Q&A entries), official documentation, and web search/fetch tools, answer ICP filing questions in one conversation: whether filing is needed, what process to follow, what rules apply, what materials are required. Consultation only, no filing operations on behalf of users.
Scope: Mainland China ICP filing (non-commercial internet information service filing) only. Not for public security filing (公安备案) or commercial filing (经营性备案). Target User: Huawei Cloud after-sales technical engineers. Site: Default China site (support.huaweicloud.com).
User Question
│
├─► catalog.yml (intent routing)
│ │
│ ▼
│ entry_point + ontology_entities
│ │
├─► knowledge-base.md
│ │
│ ▼ (if insufficient)
├─► doc-commands.md → web_fetcher.py (direct document fetch)
│ │
│ ▼ (if category unclear)
└─► web_searcher.py (search fallback, max 1 call)
This skill uses Python scripts for web document retrieval, not KooCLI. Two tools are available:
Fetches document content from a known URL via requests. No browser required.
# Fetch document text content
python {skill_dir}/scripts/web_fetcher.py fetch <URL> --mode text
# Fetch with CSS selector
python {skill_dir}/scripts/web_fetcher.py fetch <URL> --mode text --selector ".doc-content"
# Extract all links from a page
python {skill_dir}/scripts/web_fetcher.py fetch <URL> --mode links
# Fetch raw HTML
python {skill_dir}/scripts/web_fetcher.py fetch <URL> --mode html
Searches Huawei Cloud documentation site using a real browser. Requires Playwright + Chromium.
# Search Huawei Cloud docs
python {skill_dir}/scripts/web_searcher.py search "ICP备案 <keywords>"
# With remote Chrome
python {skill_dir}/scripts/web_searcher.py search "ICP备案 <keywords>" --remote-chrome-host http://host:port
references/knowledge-base.mdweb_fetcher to fetch the core document URL from references/doc-commands.mdweb_searcher when category is unclearPrerequisite check: Python >= 3.9 required; Playwright required for web search
python --version
pip install requests beautifulsoup4
# Required for web search capability
pip install playwright aiohttp
playwright install chromium
If Playwright is not installed, web_search will be unavailable; web_fetch (lightweight HTTP) still works. For remote Chrome, set environment variable
REMOTE_CHROME_HOST(e.g.,ws://remote:9222). See references/prerequisites.md for full installation guide and troubleshooting.
North Star — Any assertion must be reducible to "rule × scope × document evidence"; irreducible assertions only list gaps, no conclusions.
doc-commands.md), do not search.SKILL.md defines behavior; references/catalog.yml defines intent routing (triggers → entry_point → ontology_entities); references/filing-rules.yml defines facts and rules (grain/rules/evidence_boundary); references/doc-commands.md defines document URL index and copyable command templates; references/knowledge-base.md provides 130+ standard answer entries; references/prerequisites.md defines prerequisites.
| Phase | Task | Reference | Forbidden |
|---|---|---|---|
| Lock Intent | Match triggers to entry_point, obtain ontology_entities | catalog.yml → entry_points | Do not ask all missing context one by one |
| Gather Evidence | Check embedded knowledge → direct mapping fetch → search | knowledge-base.md → doc-commands.md → web_fetcher/web_searcher | Do not invoke tools when embedded knowledge suffices |
| Deliver | Conclusion first then facts; conclusion must be reducible to rule × scope × document | This document "Response Requirements" | No command process; no transferring investigation burden |
Match triggers in catalog.yml, determine entry_point and ontology_entities.
If boundary entry_point is matched:
NonFilingQuestion → State that it is outside ICP filing scope, suggest contacting the relevant serviceVagueQuestion → Provide basic filing concepts + process entry, suggest clarifying the specific questionNeedWorkOrder → State that a work order needs to be submittedreferences/knowledge-base.md, 80%+ high-frequency questions can be answered directlyreferences/doc-commands.md, fetch via web_fetcherweb_searcher when category is unclear, keyword format: ICP备案 + question keywordsCommand templates see references/doc-commands.md.
Output directly in Markdown or plain text, using three-part structure:
相关文档:
- 文档名称:https://...
The response is the final reply to the customer, not an internal analysis report. Follow these rules:
Q: 我们有一个新的app需要做下备案,app服务在本地,这个互联网信息服务类型怎么选呢
A: 互联网信息服务类型选择"APP"即可。但APP服务在本地服务器(非云服务器)的情况下,ICP备案需使用大陆内节点的云服务器,在哪家云服务商的服务器上部署就在哪家做备案,需确认是否有华为云大陆节点的ECS资源。 相关文档:
Q: 域名在阿里和华为都做了接入备案,取消阿里一侧的接入是否影响华为云使用?
A: 只要域名在华为云仍有接入备案,取消阿里云一侧的接入不影响华为云的使用。但如果取消后该域名没有其他接入商,会变成空壳网站/APP,被管局注销前工信部系统仍可查到。 相关文档:
Q: 备案数量已达上限怎么办?
A: 个人主体最多备案5个互联网信息服务,第6个无法提交;企业主体的上限因省份而异,参考各地区管局备案要求。另外一台ECS或L实例只能备案5个网站或APP,超出需使用新的服务器。 相关文档:
Q: 不同主体备案同一个APP,图标和备注可以一样吗?
A: 不同主体可以备案同一个APP,图标可以相同,备注按实际运营情况如实填写即可。但不同主体的APP域名需不同。 相关文档:
Q: 我们的服务器在新加坡,需要备案吗?
A: 境外服务器不需要备案,也不能完成备案。只有使用中国大陆节点服务器提供互联网信息服务时才需要ICP备案。 相关文档:
Briefing-style delivery: conclusion first, facts later; only write what was found, state scope clearly.
· list or short paragraphs, no Markdown tables.| Reference | Description | Language |
|---|---|---|
references/catalog.yml | Thin routing: triggers → entry_point → ontology_entities | zh-CN |
references/filing-rules.yml | Semantic ontology: 8 category fact definitions, core rules, evidence boundaries | zh-CN |
references/doc-commands.md | Document URL index + copyable command templates | zh-CN |
references/knowledge-base.md | 130+ standard answer knowledge base | zh-CN |
references/prerequisites.md | Prerequisites: dependency installation, environment variables, troubleshooting | en |
references/verification-method.md | Verification criteria for each workflow step and tool installation | en |