data-model-designer — Claude Code skill for construction data models
data-model-designer — a skill for Claude Code, Codex, Cursor and other agents that designs data models for construction projects: entities, relationships, schemas. Outputs SQL DDL, JSON Schema, and Mermaid ER diagrams.
MIT 116 installs ★ 434 updated 2026-10-19
npx skills add datadrivenconstruction/DDC_Skills_for_AI_Agents_in_Construction --skill data-model-designer -a claude-codeWhat data-model-designer does
The skill runs on Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, and OpenCode across macOS, Linux, and Windows. It is built in Python, so python3 must be in PATH before the skill can execute anything.
At its core are Python classes — ConstructionDataModel, Entity, Field, and Relationship. The agent defines entities with typed fields (STRING, INTEGER, FLOAT, DATE, JSON, and others), sets relationship cardinality (1:1, 1:N, N:M), and applies constraints: PRIMARY KEY, FOREIGN KEY, UNIQUE, NOT NULL. From the finished model it generates SQL DDL for PostgreSQL or other databases, a JSON Schema representation, and a Mermaid ER diagram.
The skill targets fragmented construction data: disconnected records across systems, inconsistent structures, broken links between entities. No python3 in the environment means none of the generation steps will run.
What it helps with
- Generate SQL DDL for PostgreSQL from a construction project description
- Build a Mermaid ER diagram for a set of project entities
- Define JSON Schema for domain objects like work orders or contractors
- Map relationships between entities with correct
1:NandN:Mcardinality - Apply PRIMARY KEY and FOREIGN KEY constraints to an existing schema
Example prompts for the agent
- Design a data model for a construction project with entities Project, Contractor, Task, and Material — generate PostgreSQL DDL
- Build a Mermaid ER diagram for
Project,Contractor, andWorkOrderwith their relationships - Create a JSON Schema for a
WorkOrderentity with fields for date, status, and cost
How to install data-model-designer
1Skills CLI (Vercel)
npx skills add datadrivenconstruction/DDC_Skills_for_AI_Agents_in_Construction --skill data-model-designer -a claude-code2Manually: put the files in place
~/.claude/skills/data-model-designer/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/data-model-designer/
1Skills CLI (Vercel)
npx skills add datadrivenconstruction/DDC_Skills_for_AI_Agents_in_Construction --skill data-model-designer -a codex2Manually: put the files in place
.agents/skills/data-model-designer/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add datadrivenconstruction/DDC_Skills_for_AI_Agents_in_Construction --skill data-model-designer -a cursor2Manually: put the files in place
.cursor/skills/data-model-designer/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install datadrivenconstruction/DDC_Skills_for_AI_Agents_in_Construction 2_DDC_Book/2.5-Data-Modeling-Standards/data-model-designer2Manually: put the files in place
.github/skills/data-model-designer/SKILL.md1Skills CLI (Vercel)
npx skills add datadrivenconstruction/DDC_Skills_for_AI_Agents_in_Construction --skill data-model-designer -a gemini-cli2Platform CLI
gemini skills install https://github.com/datadrivenconstruction/DDC_Skills_for_AI_Agents_in_Construction --path 2_DDC_Book/2.5-Data-Modeling-Standards/data-model-designerManually: put the files in place
.devin/skills/data-model-designer/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add datadrivenconstruction/DDC_Skills_for_AI_Agents_in_Construction --skill data-model-designer -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about data-model-designer
How do I install data-model-designer in Claude Code?
Run npx skills add datadrivenconstruction/DDC_Skills_for_AI_Agents_in_Construction --skill data-model-designer -a claude-code. The skill will be available at ~/.claude/skills/data-model-designer/SKILL.md.
What does the data-model-designer skill do?
It designs data models for construction projects: creates typed entities and fields, defines relationships between them, and generates SQL DDL, JSON Schema, and Mermaid ER diagrams.
Does data-model-designer require Python?
Yes. The skill requires python3 available in PATH. Without it, none of the generation steps will execute.
Source text
SKILL.md · MIT
---
name: "data-model-designer"
description: "Design data models for construction projects. Create entity-relationship diagrams, define schemas, and generate database structures."
homepage: "https://datadrivenconstruction.io"
metadata: {"openclaw": {"emoji": "📐", "os": ["darwin", "linux", "win32"], "homepage": "https://datadrivenconstruction.io", "requires": {"bins": ["python3"]}}}
---
# Data Model Designer
## Business Case
### Problem Statement
Construction data management challenges:
- Fragmented data across systems
- Inconsistent data structures
- Missing relationships between entities
- Difficult data integration
### Solution
Systematic data model design for construction projects, defining entities, relationships, and schemas for effective data management.
## Technical Implementation
```python
from typing import Dict, Any, List, Optional
from dataclasses import dataclass, field
from enum import Enum
import json
class DataType(Enum):
STRING = "string"
INTEGER = "integer"
FLOAT = "float"
BOOLEAN = "boolean"
DATE = "date"
DATETIME = "datetime"
TEXT = "text"
JSON = "json"
class RelationType(Enum):
ONE_TO_ONE = "1:1"
ONE_TO_MANY = "1:N"
MANY_TO_MANY = "N:M"
class ConstraintType(Enum):
PRIMARY_KEY = "primary_key"
FOREIGN_KEY = "foreign_key"
UNIQUE = "unique"
NOT_NULL = "not_null"
@dataclass
class Field:
name: str
data_type: DataType
nullable: bool = True
default: Any = None
description: str = ""
constraints: List[ConstraintType] = field(default_factory=list)
@dataclass
class Entity:
name: str
description: str
fields: List[Field] = field(default_factory=list)
primary_key: str = "id"
@dataclass
class Relationship:
name: str
from_entity: str
to_entity: str
relation_type: RelationType
from_field: str
to_field: str
class ConstructionDataModel:
"""Design data models for construction projects."""
def __init__(self, project_name: str):
self.project_name = project_name
self.entities: Dict[str, Entity] = {}
self.relationships: List[Relationship] = []
def add_entity(self, entity: Entity):
"""Add entity to model."""
self.entities[entity.name] = entity
def add_relationship(self, relationship: Relationship):
"""Add relationship between entities."""
self.relationships.append(relationship)
def create_entity(self, name: str, description: str,
fields: List[Dict[str, Any]]) -> Entity:
"""Create entity from field definitions."""
entity_fields = [
Field(
name=f['name'],
data_type=DataType(f.get('type', 'string')),
nullable=f.get('nullable', True),
default=f.get('default'),
description=f.get('description', ''),
constraints=[ConstraintType(c) for c in f.get('constraints', [])]
)
for f in fields
]
entity = Entity(name=name, description=description, fields=entity_fields)
self.add_entity(entity)
return entity
def create_relationship(self, from_entity: str, to_entity: str,
relation_type: str = "1:N",
from_field: str = None) -> Relationship:
"""Create relationship between entities."""
rel = Relationship(
name=f"{from_entity}_{to_entity}",
from_entity=from_entity,
to_entity=to_entity,
relation_type=RelationType(relation_type),
from_field=from_field or f"{to_entity.lower()}_id",
to_field="id"
)
self.add_relationship(rel)
return rel
def generate_sql_schema(self, dialect: str = "postgresql") -> str:
"""Generate SQL DDL statements."""
sql = []
type_map = {
DataType.STRING: "VARCHAR(255)",
DataType.INTEGER: "INTEGER",
DataType.FLOAT: "DECIMAL(15,2)",
DataType.BOOLEAN: "BOOLEAN",
DataType.DATE: "DATE",
DataType.DATETIME: "TIMESTAMP",
DataType.TEXT: "TEXT",
DataType.JSON: "JSONB" if dialect == "postgresql" else "JSON"
}
for name, entity in self.entities.items():
columns = []
for fld in entity.fields:
col = f" {fld.name} {type_map.get(fld.data_type, 'VARCHAR(255)')}"
if not fld.nullable:
col += " NOT NULL"
if ConstraintType.PRIMARY_KEY in fld.constraints:
col += " PRIMARY KEY"
columns.append(col)
sql.append(f"CREATE TABLE {name} (\n" + ",\n".join(columns) + "\n);")
for rel in self.relationships:
sql.append(f"""ALTER TABLE {rel.from_entity}
ADD CONSTRAINT fk_{rel.name}
FOREIGN KEY ({rel.from_field}) REFERENCES {rel.to_entity}({rel.to_field});""")
return "\n\n".join(sql)
def generate_json_schema(self) -> Dict[str, Any]:
"""Generate JSON Schema representation."""
schemas = {}
for name, entity in self.entities.items():
properties = {}
required = []
for fld in entity.fields:
prop = {"description": fld.description}
if fld.data_type == DataType.STRING:
prop["type"] = "string"
elif fld.data_type == DataType.INTEGER:
prop["type"] = "integer"
elif fld.data_type == DataType.FLOAT:
prop["type"] = "number"
elif fld.data_type == DataType.BOOLEAN:
prop["type"] = "boolean"
else:
prop["type"] = "string"
properties[fld.name] = prop
if not fld.nullable:
required.append(fld.name)
schemas[name] = {
"type": "object",
"title": entity.description,
"properties": properties,
"required": required
}
return schemas
def generate_er_diagram(self) -> str:
"""Generate Mermaid ER diagram."""
lines = ["erDiagram"]
for name, entity in self.entities.items():
for fld in entity.fields[:5]:
lines.append(f" {name} {{")
lines.append(f" {fld.data_type.value} {fld.name}")
lines.append(" }")
for rel in self.relationships:
rel_symbol = {
RelationType.ONE_TO_ONE: "||--||",
RelationType.ONE_TO_MANY: "||--o{",
RelationType.MANY_TO_MANY: "}o--o{"
}.get(rel.relation_type, "||--o{")
lines.append(f" {rel.from_entity} {rel_symbol} {rel.to_entity} : \"{rel.name}\"")
return "\n".join(lines)
def validate_model(self) -> List[str]:
"""Validate data model for issues."""
issues = []
for rel in self.relationships:
if rel.from_entity not in self.entities:
issues.append(f"Missing entity: {rel.from_entity}")
if rel.to_entity not in self.entities:
issues.append(f"Missing entity: {rel.to_entity}")
for name, entity in self.entities.items():
has_pk = any(ConstraintType.PRIMARY_KEY in f.constraints for f in entity.fields)
if not has_pk:
issues.append(f"Entity '{name}' has no primary key")
return issues
class ConstructionEntities:
"""Standard construction data entities."""
@staticmethod
def project_entity() -> Entity:
return Entity(
name="projects",
description="Construction projects",
fields=[
Field("id", DataType.INTEGER, False, constraints=[ConstraintType.PRIMARY_KEY]),
Field("code", DataType.STRING, False, constraints=[ConstraintType.UNIQUE]),
Field("name", DataType.STRING, False),
Field("status", DataType.STRING),
Field("start_date", DataType.DATE),
Field("end_date", DataType.DATE),
Field("budget", DataType.FLOAT)
]
)
@staticmethod
def activity_entity() -> Entity:
return Entity(
name="activities",
description="Schedule activities",
fields=[
Field("id", DataType.INTEGER, False, constraints=[ConstraintType.PRIMARY_KEY]),
Field("project_id", DataType.INTEGER, False),
Field("wbs_code", DataType.STRING),
Field("name", DataType.STRING, False),
Field("start_date", DataType.DATE),
Field("end_date", DataType.DATE),
Field("percent_complete", DataType.FLOAT)
]
)
@staticmethod
def cost_item_entity() -> Entity:
return Entity(
name="cost_items",
description="Project cost items",
fields=[
Field("id", DataType.INTEGER, False, constraints=[ConstraintType.PRIMARY_KEY]),
Field("project_id", DataType.INTEGER, False),
Field("wbs_code", DataType.STRING),
Field("description", DataType.STRING),
Field("budgeted_cost", DataType.FLOAT),
Field("actual_cost", DataType.FLOAT)
]
)
```
## Quick Start
```python
# Create model
model = ConstructionDataModel("Office Building A")
# Add standard entities
model.add_entity(ConstructionEntities.project_entity())
model.add_entity(ConstructionEntities.activity_entity())
model.add_entity(ConstructionEntities.cost_item_entity())
# Add relationships
model.create_relationship("activities", "projects")
model.create_relationship("cost_items", "projects")
# Generate SQL
sql = model.generate_sql_schema("postgresql")
print(sql)
```
## Common Use Cases
### 1. Custom Entity
```python
model.create_entity(
name="change_orders",
description="Project change orders",
fields=[
{"name": "id", "type": "integer", "nullable": False, "constraints": ["primary_key"]},
{"name": "project_id", "type": "integer", "nullable": False},
{"name": "amount", "type": "float"},
{"name": "status", "type": "string"}
]
)
```
### 2. Generate ER Diagram
```python
er_diagram = model.generate_er_diagram()
print(er_diagram)
```
### 3. Validate Model
```python
issues = model.validate_model()
for issue in issues:
print(f"Issue: {issue}")
```
## Resources
- **DDC Book**: Chapter 2.5 - Data Models and Standards
- **Website**: https://datadrivenconstruction.io
© datadrivenconstruction · MIT · datadrivenconstruction/DDC_Skills_for_AI_Agents_in_Construction