You've already forked VARAM_classification_schema
Import UAPF package
20 lines
1.8 KiB
YAML
MCP Server
20 lines
1.8 KiB
YAML
MCP Server
# ProcessGit MCP Configuration
|
|
# Enables MCP endpoint at: https://processgit.org/VARAM/Document_classification_schema/mcp
|
|
# Trigger file — ProcessGit auto-detects this in repository root
|
|
|
|
version: 1
|
|
|
|
server:
|
|
name: "VARAM Document Classification Schema"
|
|
description: "Machine-readable classification schema for VARAM DVS Namejs document management system. Functional categories (V2) with AI context descriptions (IETVER/NEIETVER), programme/project/audience metadata, restructured per MK noteikumi Nr. 282 §33. P-/I1-/I2- prefix numbering preserved from V1."
|
|
instructions: "Query document classification categories with AI-friendly context descriptions. Use 'search' to find categories by name, code, or description keywords — each category contains structured IETVER (includes) and NEIETVER (excludes with cross-reference codes) sections enabling precise AI document classification. Use 'list_entities' to browse by type (category, department, programme, project, audience), 'get_entity' for full details by ID (e.g. category:P-1-13, project:PRJ-005, department:IPD), 'generate_document' for formatted classification table. For AI classification workflow: search by document content keywords, then compare IETVER/NEIETVER in returned descriptions to select best match."
|
|
|
|
sources:
|
|
- path: "classification-v2-ai.xml"
|
|
type: "xml"
|
|
description: "V2 functional classification schema — 159 categories with AI context descriptions (IETVER/NEIETVER cross-references), 22 groups, 3 domains (P=Pārvalde, I1=Investīciju programmu vadība, I2=Investīciju projektu ieviešana), conditional retention rules, programme/project/audience metadata tags"
|
|
|
|
- path: "vocabularies.xml"
|
|
type: "xml"
|
|
description: "Controlled vocabularies — 22 departments, 13 programmes (PRG-xxx), 33 projects (PRJ-xxx), 9 audience types"
|