GEO Structured Knowledge Base Generator

Problem

AI search and answer engines often lack a stable, citable source when asked what a company is, what products it offers, whether it is credible, and how to contact it. Loose web links are hard to retrieve and maintain. This skill turns company information into a structured knowledge base oriented toward GEO brand retrieval.

How It Works

It executes a four-step workflow through IMA OpenAPI:
- Parameter extraction: identifies company name and industry from user input, asking at most one follow-up round if details are missing.
- Knowledge base creation: calls create_knowledge_base for shared, personal, or subscribed bases; skips this step when an existing kb_id is provided.
- Folder creation: creates 10 folders from references/folder-spec.md, such as Company Profile, Product System, Customer Cases, Certifications, FAQ, and Contact & Locations.
- Structured reporting: returns creation results, kb_id, folder_id, and industry-specific recommended questions.

The 10 folders are not arbitrary categories. They cover the retrieval path of identity, product, trust, traffic, and conversion, making later content easier for AI systems to cite consistently.

Boundaries And Notes

• The default plan creates 10 folders. They can be adjusted, but writes are irreversible, so confirm the company name and kb_id before execution.
• Recommended questions must be industry-specific; generic templates are not acceptable.
• Numeric MCP knowledge-base IDs cannot be used directly with OpenAPI writes. Search by name first to obtain the OpenAPI encrypted string ID.
• For Chinese payloads on Windows, avoid curl -d with manually concatenated JSON. Use Node.js to send requests and prevent encoding corruption.