kb-mcp-server
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| KB_TENANT_ID | Yes | Tenant ID (UUID) from Supabase auth.users | |
| SUPABASE_URL | Yes | Supabase project URL | |
| SILICONFLOW_API_KEY | Yes | SiliconFlow API key for embeddings | |
| SUPABASE_SERVICE_ROLE_KEY | Yes | Supabase service role key |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| search_knowledge_baseA | 语义检索知识库,返回与查询最相关的若干片段(含相似度与所属文档标题)。 |
| list_documentsA | 列出当前知识库的全部文档及元数据(id、标题、类型、状态、片段数、创建时间)。 |
| get_document_contentA | 按 document_id 拉取并拼接单篇文档全文。 |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
| kb_qa | 知识库问答提示词模板:把用户原始问题包装成"先检索后作答 + 标注来源"的指令。 用法:在 Desktop "Prompts" 面板选择 kb_qa,填入 question,即发送为一条用户消息。 |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| documents_snapshot | 知识库文档清单的只读快照(JSON 字符串)。 与 list_documents tool 的区别: - Resource:**客户端主动读取**的上下文数据,Desktop 用户可在"附件/资源"面板里 手动挂载;模型不会自动调用,适合作为对话起手的背景资料 - Tool:**模型决策调用**,模型在需要时自行触发(如用户问"知识库有什么") 返回与 list_documents 同结构的 JSON;空清单返回中文文案。 |
TDQS
Scored across 3 tools
Each tool has a distinct purpose: listing documents, searching semantically, and retrieving full content. No ambiguity between them; descriptions clearly differentiate when to use each.
All tool names follow the consistent verb_noun pattern in snake_case (list_documents, search_knowledge_base, get_document_content), making the API predictable.
Three tools cover the core operations for a read-only knowledge base server: list, search, and get content. The count is well-scoped with no redundancy.
The set provides essential read operations, but lacks create/update/delete actions. For a purely retrieval-focused knowledge base, this is nearly complete; minor gap for full CRUD.