MCPInstall

bim-validation-pipeline

Build automated BIM validation pipelines for IFC/Revit data. Continuous validation against IDS, LOD requirements, COBie, and project-specific BEP standards.

Installs
122
Last 24h
—
Trending rank
—
Security
Not audited

Install

Works with every agent the skills CLI supports.

npx skills add https://github.com/datadrivenconstruction/ddc_skills_for_ai_agents_in_construction --skill bim-validation-pipeline
  1. 01

    Install

    Run the command in your project. The CLI asks which agents to install for.

  2. 02

    Describe the task

    Ask your agent for something this skill covers. It loads bim-validation-pipeline automatically when relevant.

  3. 03

    Review the result

    The agent follows the SKILL.md steps, scripts and references bundled with the skill.

SKILL.md

The instructions your agent reads when the skill is triggered.

BIM Validation Pipeline

Overview

Based on DDC methodology (Chapter 4.3), this skill provides automated BIM data validation pipelines. Validate BIM models against Information Delivery Specification (IDS), Level of Development (LOD) requirements, and project standards.

Book Reference: "Автоматический ETL конвейер для валидации данных" / "Automated ETL Pipeline for Data Validation"

"Автоматизированная валидация BIM-данных позволяет выявлять ошибки на ранних стадиях и обеспечивать соответствие требованиям BEP." — DDC Book, Chapter 4.3

Quick Start

import ifcopenshell
import pandas as pd

# Load IFC model
ifc_model = ifcopenshell.open("model.ifc")

# Quick validation checks
walls = ifc_model.by_type("IfcWall")
print(f"Total walls: {len(walls)}")

# Check for required properties
issues = []
for wall in walls:
    # Check if wall has material
    if not wall.HasAssociations:
        issues.append(f"Wall {wall.GlobalId}: No material assigned")

print(f"Issues found: {len(issues)}")

BIM Validation Framework

Core Validator Class

import ifcopenshell
import ifcopenshell.util.element as element_util
import pandas as pd
from dataclasses import dataclass
from typing import List, Dict, Optional
from enum import Enum

class Severity(Enum):
    ERROR = "error"
    WARNING = "warning"
    INFO = "info"

@dataclass
class ValidationIssue:
    element_id: str
    element_type: str
    rule_id: str
    severity: Severity
    message: str
    location: Optional[str] = None

class BIMValidator:
    """Comprehensive BIM model validator"""

    def __init__(self, ifc_path: str):
        self.model = ifcopenshell.open(ifc_path)
        self.issues: List[ValidationIssue] = []
        self.stats = {}

    def validate_all(self):
        """Run all validation checks"""
        self.validate_geometry()
        self.validate_properties()
        self.validate_relationships()
        self.validate_naming()
        self.validate_classification()
        return self.get_report()

    def validate_geometry(self):
        """Check geometry validity"""
        elements_with_geometry = [
            e for e in self.model.by_type("IfcProduct")
            if e.Representation
        ]

        for element in elements_with_geometry:
            # Check for zero volume
            try:
                settings = ifcopenshell.geom.settings()
                shape = ifcopenshell.geom.create_shape(settings, element)
                # Volume check would go here
            except:
                self.issues.append(ValidationIssue(
                    element_id=element.GlobalId,
                    element_type=element.is_a(),
                    rule_id="GEO-001",
                    severity=Severity.ERROR,
                    message="Invalid or missing geometry"
                ))

        self.stats['elements_with_geometry'] = len(elements_with_geometry)

    def validate_properties(self, required_psets: Dict[str, List[str]] = None):
        """Check required property sets and properties"""
        if required_psets is None:
            required_psets = {
                'IfcWall': ['Pset_WallCommon', 'BaseQuantities'],
                'IfcSlab': ['Pset_SlabCommon', 'BaseQuantities'],
                'IfcColumn': ['Pset_ColumnCommon', 'BaseQuantities'],
                'IfcBeam': ['Pset_BeamCommon', 'BaseQuantities']
            }

        for ifc_type, psets in required_psets.items():
            elements = self.model.by_type(ifc_type)

            for element in elements:
                element_psets = element_util.get_psets(element)

                for required_pset in psets:
                    if required_pset not in element_psets:
                        self.issues.append(ValidationIssue(
                            element_id=element.GlobalId,
                            element_type=ifc_type,
                            rule_id="PROP-001",
                            severity=Severity.WARNING,
                            message=f"Missing PropertySet: {required_pset}"
                        ))

    def validate_relationships(self):
        """Check spatial containment and relationships"""
        products = self.model.by_type("IfcProduct")

        for product in products:
            # Check spatial containment
            if hasattr(product, 'ContainedInStructure'):
                if not product.ContainedInStructure:
                    self.issues.append(ValidationIssue(
                        element_id=product.GlobalId,
                        element_type=product.is_a(),
                        rule_id="REL-001",
                        severity=Severity.WARNING,
                        message="Element not assigned to building storey"
                    ))

            # Check material assignment
            if hasattr(product, 'HasAssociations'):
                has_material = any(
                    rel.is_a('IfcRelAssociatesMaterial')
                    for rel in (product.HasAssociations or [])
                )
                if not has_material and product.is_a() in ['IfcWall', 'IfcSlab', 'IfcColumn']:
                    self.issues.append(ValidationIssue(
                        element_id=product.GlobalId,
                        element_type=product.is_a(),
                        rule_id="MAT-001",
                        severity=Severity.WARNING,
                        message="No material assigned"
                    ))

    def validate_naming(self, patterns: Dict[str, str] = None):
        """Validate element naming conventions"""
        import re

        if patterns is None:
            patterns = {
                'IfcBuildingStorey': r'^(Level|L|Floor|Уровень)\s*[-]?\d+',
                'IfcWall': r'^W[-_]?\d{3,}|^Wall[-_]',
                'IfcColumn': r'^C[-_]?\d{3,}|^Column[-_]',
                'IfcSpace': r'^Room[-_]|^Space[-_]'
            }

        for ifc_type, pattern in patterns.items():
            elements = self.model.by_type(ifc_type)

            for element in elements:
                name = element.Name or ""
                if not re.match(pattern, name, re.IGNORECASE):
                    self.issues.append(ValidationIssue(
                        element_id=element.GlobalId,
                        element_type=ifc_type,
                        rule_id="NAME-001",
                        severity=Severity.INFO,
                        message=f"Name '{name}' doesn't match convention"
                    ))

    def validate_classification(self, required_systems: List[str] = None):
        """Check classification system assignments"""
        if required_systems is None:
            required_systems = ['Uniclass', 'OmniClass', 'Uniformat']

        elements = self.model.by_type("IfcProduct")

        for element in elements:
            if hasattr(element, 'HasAssociations'):
                has_classification = any(
                    rel.is_a('IfcRelAssociatesClassification')
                    for rel in (element.HasAssociations or [])
                )

                if not has_classification:
                    self.issues.append(ValidationIssue(
                        element_id=element.GlobalId,
                        element_type=element.is_a(),
                        rule_id="CLASS-001",
                        severity=Severity.INFO,
                        message="No classification assigned"
                    ))

    def get_report(self):
        """Generate validation report"""
        by_severity = {s: [] for s in Severity}
        by_type = {}
        by_rule = {}

        for issue in self.issues:
            by_severity[issue.severity].append(issue)

            if issue.element_type not in by_type:
                by_type[issue.element_type] = []
            by_type[issue.element_type].append(issue)

            if issue.rule_id not in by_rule:
                by_rule[issue.rule_id] = []
            by_rule[issue.rule_id].append(issue)

        return {
            'total_issues': len(self.issues),
            'errors': len(by_severity[Severity.ERROR]),
            'warnings': len(by_severity[Severity.WARNING]),
            'info': len(by_severity[Severity.INFO]),
            'by_type': {k: len(v) for k, v in by_type.items()},
            'by_rule': {k: len(v) for k, v in by_rule.items()},
            'issues': self.issues,
            'stats': self.stats
        }

LOD Validation

Level of Development Checker

class LODValidator:
    """Validate Level of Development (LOD) requirements"""

    # LOD requirements by element type
    LOD_REQUIREMENTS = {
        'LOD100': {
            'geometry': False,
            'properties': [],
            'description': 'Conceptual'
        },
        'LOD200': {
            'geometry': True,
            'approximate_size': True,
            'properties': ['Category'],
            'description': 'Schematic Design'
        },
        'LOD300': {
            'geometry': True,
            'exact_size': True,
            'properties': ['Category', 'Material', 'Type'],
            'quantities': ['Length', 'Area', 'Volume'],
            'description': 'Design Development'
        },
        'LOD350': {
            'geometry': True,
            'exact_size': True,
            'properties': ['Category', 'Material', 'Type', 'Manufacturer'],
            'quantities': ['Length', 'Area', 'Volume', 'Weight'],
            'connections': True,
            'description': 'Construction Documentation'
        },
        'LOD400': {
            'geometry': True,
            'fabrication_ready': True,
            'properties': ['Category', 'Material', 'Type', 'Manufacturer',
                          'Model', 'Serial', 'InstallationDate'],
            'quantities': ['All'],
            'connections': True,
            'description': 'Fabrication & Assembly'
        }
    }

    def __init__(self, model, target_lod='LOD300'):
        self.model = model
        self.target_lod = target_lod
        self.requirements = self.LOD_REQUIREMENTS.get(target_lod, {})
        self.results = []

    def validate_element(self, element):
        """Validate single element against LOD requirements"""
        issues = []
        element_guid = element.GlobalId
        psets = element_util.get_psets(element)

        # Check geometry
        if self.requirements.get('geometry'):
            if not element.Representation:
                issues.append({
                    'element': element_guid,
                    'issue': 'Missing geometry',
                    'required_for': self.target_lod
                })

        # Check required properties
        required_props = self.requirements.get('properties', [])
        all_props = {}
        for pset_name, props in psets.items():
            all_props.update(props)

        for prop in required_props:
            if prop not in all_props or all_props[prop] is None:
                issues.append({
                    'element': element_guid,
                    'issue': f'Missing property: {prop}',
                    'required_for': self.target_lod
                })

        # Check quantities
        required_quantities = self.requirements.get('quantities', [])
        if required_quantities != ['All']:
            qsets = psets.get('BaseQuantities', {})
            for qty in required_quantities:
                if qty not in qsets:
                    issues.append({
                        'element': element_guid,
                        'issue': f'Missing quantity: {qty}',
                        'required_for': self.target_lod
                    })

        return issues

    def validate_model(self, element_types=None):
        """Validate entire model"""
        if element_types is None:
            element_types = ['IfcWall', 'IfcSlab', 'IfcColumn', 'IfcBeam',
                            'IfcDoor', 'IfcWindow', 'IfcStair']

        all_issues = []
        summary = {}

        for ifc_type in element_types:
            elements = self.model.by_type(ifc_type)
            type_issues = []

            for element in elements:
                issues = self.validate_element(element)
                type_issues.extend(issues)

            summary[ifc_type] = {
                'total': len(elements),
                'issues': len(type_issues),
                'compliance': ((len(elements) - len(type_issues)) /
                              len(elements) * 100) if elements else 100
            }
            all_issues.extend(type_issues)

        return {
            'target_lod': self.target_lod,
            'total_issues': len(all_issues),
            'summary': summary,
            'issues': all_issues
        }

IDS Validation

Information Delivery Specification

import xml.etree.ElementTree as ET

class IDSValidator:
    """Validate against IDS (Information Delivery Specification)"""

    def __init__(self, ids_path: str):
        self.ids = self._parse_ids(ids_path)

    def _parse_ids(self, path):
        """Parse IDS XML file"""
        tree = ET.parse(path)
        root = tree.getroot()

        specifications = []
        for spec in root.findall('.//specification'):
            specifications.append({
                'name': spec.get('name'),
                'applicability': self._parse_facets(spec.find('applicability')),
                'requirements': self._parse_facets(spec.find('requirements'))
            })

        return specifications

    def _parse_facets(self, element):
        """Parse IDS facets"""
        if element is None:
            return []

        facets = []
        for child in element:
            facet = {
                'type': child.tag,
                'constraints': {}
            }
            for attr, value in child.attrib.items():
                facet['constraints'][attr] = value
            facets.append(facet)

        return facets

    def validate(self, model):
        """Validate IFC model against IDS"""
        results = []

        for spec in self.ids:
            applicable_elements = self._find_applicable_elements(
                model, spec['applicability']
            )

            for element in applicable_elements:
                issues = self._check_requirements(element, spec['requirements'])
                if issues:
                    results.append({
                        'specification': spec['name'],
                        'element': element.GlobalId,
                        'issues': issues
                    })

        return results

    def _find_applicable_elements(self, model, applicability):
        """Find elements matching applicability criteria"""
        elements = []

        for facet in applicability:
            if facet['type'] == 'entity':
                ifc_type = facet['constraints'].get('name')
                if ifc_type:
                    elements.extend(model.by_type(ifc_type))

        return elements

    def _check_requirements(self, element, requirements):
        """Check element against requirements"""
        issues = []
        psets = element_util.get_psets(element)

        for req in requirements:
            if req['type'] == 'property':
                pset_name = req['constraints'].get('propertySet')
                prop_name = req['constraints'].get('name')

                if pset_name and prop_name:
                    pset = psets.get(pset_name, {})
                    if prop_name not in pset:
                        issues.append(f"Missing property: {pset_name}.{prop_name}")

        return issues

Pipeline Automation

Automated Validation Pipeline

import os
from datetime import datetime
import json

class BIMValidationPipeline:
    """Automated BIM validation pipeline"""

    def __init__(self, config_path=None):
        self.config = self._load_config(config_path)
        self.results_history = []

    def _load_config(self, path):
        if path and os.path.exists(path):
            with open(path) as f:
                return json.load(f)

        return {
            'lod_target': 'LOD300',
            'required_psets': {
                'IfcWall': ['Pset_WallCommon'],
                'IfcSlab': ['Pset_SlabCommon']
            },
            'naming_patterns': {},
            'fail_on_errors': True,
            'warn_threshold': 50
        }

    def run(self, ifc_path):
        """Run complete validation pipeline"""
        start_time = datetime.now()

        # Initialize validators
        validator = BIMValidator(ifc_path)
        lod_validator = LODValidator(
            validator.model,
            self.config['lod_target']
        )

        # Run validations
        bim_report = validator.validate_all()
        lod_report = lod_validator.validate_model()

        # Compile results
        result = {
            'file': ifc_path,
            'timestamp': start_time.isoformat(),
            'duration_seconds': (datetime.now() - start_time).total_seconds(),
            'bim_validation': bim_report,
            'lod_validation': lod_report,
            'passed': self._evaluate_pass(bim_report, lod_report)
        }

        self.results_history.append(result)
        return result

    def _evaluate_pass(self, bim_report, lod_report):
        """Determine if validation passed"""
        if self.config['fail_on_errors'] and bim_report['errors'] > 0:
            return False

        if bim_report['warnings'] > self.config['warn_threshold']:
            return False

        return True

    def run_batch(self, ifc_paths):
        """Run validation on multiple files"""
        results = []
        for path in ifc_paths:
            try:
                result = self.run(path)
                results.append(result)
            except Exception as e:
                results.append({
                    'file': path,
                    'error': str(e),
                    'passed': False
                })

        return {
            'total': len(results),
            'passed': sum(1 for r in results if r.get('passed', False)),
            'failed': sum(1 for r in results if not r.get('passed', True)),
            'results': results
        }

    def export_report(self, output_path):
        """Export validation results to Excel"""
        if not self.results_history:
            return None

        latest = self.results_history[-1]

        with pd.ExcelWriter(output_path, engine='openpyxl') as writer:
            # Summary
            summary = pd.DataFrame({
                'Metric': ['File', 'Timestamp', 'Passed', 'Errors', 'Warnings'],
                'Value': [
                    latest['file'],
                    latest['timestamp'],
                    latest['passed'],
                    latest['bim_validation']['errors'],
                    latest['bim_validation']['warnings']
                ]
            })
            summary.to_excel(writer, sheet_name='Summary', index=False)

            # Issues
            if latest['bim_validation']['issues']:
                issues_df = pd.DataFrame([
                    {
                        'Element': i.element_id,
                        'Type': i.element_type,
                        'Rule': i.rule_id,
                        'Severity': i.severity.value,
                        'Message': i.message
                    }
                    for i in latest['bim_validation']['issues']
                ])
                issues_df.to_excel(writer, sheet_name='Issues', index=False)

        return output_path

Quick Reference

Rule IDDescriptionSeverity
GEO-001Invalid/missing geometryERROR
PROP-001Missing PropertySetWARNING
REL-001No spatial containmentWARNING
MAT-001No material assignedWARNING
NAME-001Invalid naming conventionINFO
CLASS-001No classificationINFO

LOD Requirements Summary

LODGeometryPropertiesQuantities
100No--
200ApproximateCategory-
300ExactMaterial, TypeL, A, V
350Exact + connectionsManufacturerAll
400Fabrication-readyAll detailsAll

Resources

Next Steps

  • See ifc-data-extraction for extracting data from IFC
  • See data-quality-check for general data validation
  • See qto-report for quantity take-off from validated models

FAQ

What does the bim-validation-pipeline skill do?

Build automated BIM validation pipelines for IFC/Revit data. Continuous validation against IDS, LOD requirements, COBie, and project-specific BEP standards.

How do I install bim-validation-pipeline?

Run npx skills add https://github.com/datadrivenconstruction/ddc_skills_for_ai_agents_in_construction --skill bim-validation-pipeline in your project. The skills CLI detects the agents you use and copies the skill into the right folder. For Claude Code you can also place the skill folder at ~/.claude/skills/bim-validation-pipeline/ (personal) or .claude/skills/bim-validation-pipeline/ (project).

Which AI agents can use bim-validation-pipeline?

Any agent that supports the open Agent Skills format (a folder with a SKILL.md file), including Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Windsurf, Cline, OpenCode and more.

Is bim-validation-pipeline safe to install?

No partner security audit is available yet. Skills can run commands and scripts, so read the SKILL.md and any bundled files before installing.

Who maintains bim-validation-pipeline?

It is published by datadrivenconstruction in datadrivenconstruction/ddc_skills_for_ai_agents_in_construction. It has 122 recorded installs.

More from datadrivenconstruction/ddc_skills_for_ai_agents_in_construction

View all →

warranty-tracker

datadrivenconstruction/ddc_skills_for_ai_agents_in_construction

Track and manage construction warranties. Monitor expiration dates, claims, and manufacturer documentation.

AI & Agents996 installs

cad-to-data

datadrivenconstruction/ddc_skills_for_ai_agents_in_construction

Convert CAD/BIM files to structured data. Extract element data from Revit, IFC, DWG, DGN files.

AI & Agents333 installs

dwg-to-excel

datadrivenconstruction/ddc_skills_for_ai_agents_in_construction

Open the skill to read its SKILL.md instructions.

Data & Databases309 installs

drawing-analyzer

datadrivenconstruction/ddc_skills_for_ai_agents_in_construction

Open the skill to read its SKILL.md instructions.

AI & Agents259 installs

permit-tracking-automation

datadrivenconstruction/ddc_skills_for_ai_agents_in_construction

Automate construction permit tracking and management. Monitor application status, track renewal deadlines, manage document requirements, and integrate with municipal systems.

cost-estimation-resource

datadrivenconstruction/ddc_skills_for_ai_agents_in_construction

Calculate construction costs using resource-based method. Estimate project costs from work items, physical resource norms, and current prices.

AI & Agents193 installs

Popular ai & agents skills

View category →

google-agents-cli-publish

google/agents-cli

This skill should be used when the user wants to "publish an agent", "publish my ADK agent", "register an agent with Gemini", "publish to Gemini" (formerly Gemini Enterprise), or needs guidance on the agents-cli publish gemini command. Also use when the user wants to "manage agents in Agent Registry", "list/update/delete registered agents", or "register an MCP server". Covers ADK vs A2A registration modes, programmatic and interactive usage, flag reference, auto-detection from deployment metadata, Agent Registry fleet management, and troubleshooting. Part of the agents-cli skills suite. Do NOT use for deployment (use google-agents-cli-deploy).

AI & Agents513.9K installs

google-agents-cli-eval

google/agents-cli

This skill should be used when the user wants to "run an evaluation", "evaluate my agent", "evaluate my ADK agent", "write an eval dataset", "analyze eval failures", "compare eval results", "optimize agent", or needs guidance on the Gemini platform eval methodology and the Quality Flywheel. Covers eval metrics, dataset schema, LLM-as-judge scoring, and common failure causes. Applies to any agents-cli project, whatever framework the agent is written in. Do NOT use for agent API code patterns (ADK: use google-agents-cli-adk-code), deployment (use google-agents-cli-deploy), or project scaffolding (use google-agents-cli-scaffold).

AI & Agents513.5K installs

skill-creator

anthropics/skills

Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.

OfficialAI & Agents416.4K installs

codex-pet

prime-skills/runcomfy-agent-skills

Codex Pet generator on RunComfy. Build a Codex-compatible Codex Pet spritesheet.webp + pet.json from a single reference image, drop it into ${CODEXHOME:-$HOME/.codex}/pets/ / and Codex picks it up as a custom Codex Pet next to the 8 built-ins. This skill produces the exact Codex Pet atlas Codex expects (1536x1872 PNG/WebP, 8 cols x 9 rows, 192x208 cells, 9 animation states — idle, running-right, running-left, waving, jumping, failed, waiting, running, review). Calls OpenAI GPT Image 2 edit ONCE via the local RunComfy CLI as runcomfy run openai/gpt-image-2/edit to produce a canonical Codex Pet pose, then assembles all 9 animation rows programmatically with ImageMagick micro-transforms — no Codex Pro, no $imagegen, no OPENAIAPIKEY required, only RUNCOMFYTOKEN. Triggers on "codex pet", "create codex pet", "make codex pet", "hatch codex pet", "/hatch image", "desktop pet codex", "codex pets", "spritesheet.webp", or any explicit ask to build a custom pet for OpenAI Codex.

AI & Agents402.1K installs

full-output-enforcement

leonxlnx/taste-skill

Overrides default LLM truncation behavior. Enforces complete code generation, bans placeholder patterns, and handles token-limit splits cleanly. Apply to any task requiring exhaustive, unabridged output.

AI & Agents401K installs

executing-plans

obra/superpowers

Use when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available

AI & Agents245.7K installs