Overview

The /czech-autocrawler-supreme command orchestrates autonomous business intelligence gathering from Czech government registries using a 3NL (3-Layer Neural Learning) architecture with AIAD agent coordination. Czech Republic business registries -- ARES (Administrative Register of Economic Subjects), Justice.cz (Commercial Register), and CUZK (Czech Office for Surveying, Mapping and Cadastre) -- contain authoritative data on corporate entities, ownership structures, property holdings, and business relationships. This command automates the systematic extraction, cross-referencing, and indexing of that intelligence.

The 3NL architecture provides three processing layers: L1 Pattern Recognition handles Czech-specific data validation, HTML parsing, and quality assessment; L2 Knowledge Synthesis performs entity resolution, deduplication, and relationship mapping across registries; L3 Meta-Learning optimizes crawling strategies, predicts performance, and detects data anomalies. Together, these layers transform raw registry data into structured, searchable business intelligence with measurable quality metrics.

This command operates under the L3 authority level and is executed by the czech-autocrawler agent. It is part of the platform's 216-command slash command registry, built on the AIAD (Autonomous Intelligence Agent Design) standard. The command coordinates multiple specialized agents across all three neural layers and integrates with Meilisearch for real-time search indexing of discovered entities.

Business intelligence gathered through this command feeds into the broader OSINT capability of the Prismatic Platform, providing corporate structure analysis, beneficial ownership investigations, and regulatory compliance verification. All operations comply with GDPR data protection requirements and respect each registry's terms of service and rate limiting policies.

Architecture

3NL Processing Pipeline

CZECH REGISTRY SOURCES
======================

ARES (ICO/DICO lookup)  +  Justice.cz (Company records)  +  CUZK (Cadastral data)
         |                           |                            |
         v                           v                            v
    +------------------------------------------------------------+
    |           L1: PATTERN RECOGNITION LAYER                     |
    |  +-- Czech identifier validation (ICO, DICO, RC)           |
    |  +-- HTML/XML parsing and data extraction                   |
    |  +-- Quality assessment and confidence scoring              |
    |  +-- Rate limit management and request scheduling           |
    +------------------------------------------------------------+
                              |
                              v
    +------------------------------------------------------------+
    |           L2: KNOWLEDGE SYNTHESIS LAYER                     |
    |  +-- Cross-registry entity resolution                      |
    |  +-- Deduplication with fuzzy matching                      |
    |  +-- Relationship mapping (ownership, management)          |
    |  +-- Data enrichment and gap filling                       |
    +------------------------------------------------------------+
                              |
                              v
    +------------------------------------------------------------+
    |           L3: META-LEARNING LAYER                           |
    |  +-- Crawling strategy optimization                        |
    |  +-- Performance prediction and resource planning          |
    |  +-- Anomaly detection in entity data                      |
    |  +-- Quality threshold adaptive adjustment                 |
    +------------------------------------------------------------+
                              |
                              v
    +------------------------------------------------------------+
    |           MEILISEARCH INDEX                                 |
    |  +-- Real-time entity indexing                              |
    |  +-- Full-text search across all fields                    |
    |  +-- Faceted search (region, industry, size)               |
    |  +-- Relationship-based queries                            |
    +------------------------------------------------------------+

Agent Coordination

Agent	Layer	Role
`czech-autocrawler-supreme`	Supreme	Orchestrator coordinating all layers
`l1-pattern-engine`	L1	Czech data validation and extraction
`l2-entity-resolver`	L2	Cross-registry entity resolution
`l3-strategy-optimizer`	L3	Crawling strategy and performance optimization

Usage

Discovery Mode (Default)

# Start discovery mode with default settings
# Sources: all registries, batch size: 100, real-time indexing, 80% quality threshold
/czech-autocrawler-supreme

# Discovery focused on specific registry
/czech-autocrawler-supreme --mode discovery --sources ares

# Discovery with custom batch size
/czech-autocrawler-supreme --mode discovery --batch-size 200

Enrichment Mode

# Enrich existing entities from ARES and Justice.cz
/czech-autocrawler-supreme --mode enrichment --sources ares,justice --batch-size 50

# Targeted enrichment of specific entity
/czech-autocrawler-supreme --mode enrichment --target ICO:12345678

Validation Mode

# Validate data quality with strict threshold
/czech-autocrawler-supreme --mode validation --quality-threshold 0.9

# Validation with comprehensive reporting
/czech-autocrawler-supreme --mode validation --quality-threshold 0.85 --verbose

Full Processing Mode

# Full pipeline: discovery + enrichment + validation
/czech-autocrawler-supreme --mode full --batch-size 250 --indexing batch

# Full mode with CUZK focus for real estate intelligence
/czech-autocrawler-supreme --mode full --sources cuzk --quality-threshold 0.75

Options & Parameters

Parameter	Type	Required	Default	Description
--mode	enum	No	`discovery`	Operation mode: discovery, enrichment, validation, full
--sources	array	No	`ares,justice,cuzk`	Registry sources to crawl
--batch-size	integer	No	100	Entities per batch (1-500)
--indexing	enum	No	`realtime`	Indexing mode: realtime, batch, disabled
--quality-threshold	float	No	0.8	Minimum quality score (0.0-1.0)
--target	string	No	--	Specific entity identifier (ICO:XXXXXXXX)
--verbose	boolean	No	false	Show detailed processing output

Execution Flow

/czech-autocrawler-supreme [options]
    |
    v
PHASE 1: INITIALIZATION (< 5s)
    +-- Activate czech-autocrawler-supreme agent
    +-- Initialize 3NL layers (L1, L2, L3)
    +-- Establish registry connections
    +-- Configure Meilisearch index and schema
    +-- Load L3 strategy optimization state
    |
    v
PHASE 2: DISCOVERY & PROCESSING (variable)
    +-- L1: Pattern recognition across registries
    +-- L1: Czech identifier validation (ICO, DICO)
    +-- L2: Cross-registry entity resolution
    +-- L2: Relationship mapping (ownership chains)
    +-- L3: Strategy optimization based on throughput
    +-- L3: Performance prediction and adjustment
    |
    v
PHASE 3: INDEXING & REPORTING (< 10s per batch)
    +-- Stream processed entities to Meilisearch
    +-- Generate quality assessment reports
    +-- Track throughput and error metrics
    +-- Update GitLab milestone progress
    |
    v
PHASE 4: COMPLETION
    +-- Final quality metrics summary
    +-- Relationship graph statistics
    +-- Performance analytics report
    +-- Next crawl recommendations

Output Structure

{
  "status": "success",
  "crawl_id": "czreg_1703347200_a1b2c3d4",
  "summary": {
    "entities_discovered": 1250,
    "entities_processed": 1200,
    "entities_indexed": 1195,
    "quality_score": 0.92,
    "processing_time_ms": 45000,
    "registries_used": ["ares", "justice", "cuzk"]
  },
  "quality_metrics": {
    "completeness": 0.95,
    "accuracy": 0.94,
    "consistency": 0.89,
    "freshness": 0.98
  },
  "relationships": {
    "ownership_chains": 45,
    "management_links": 128,
    "business_partnerships": 67
  }
}

Integration Points

Component	Relationship	Details
Prismatic Agents	Multi-agent coordination	4 agents across 3NL layers
AIAD Registry	Command specification	Supreme coordination category
Meilisearch	Real-time search indexing	`czech_business_entities` index
Quality Gates	Data quality enforcement	Quality threshold gating
Telemetry	Performance metrics	Throughput and error rate tracking
GitLab Milestones	Project management	Milestone MS-6250279 progress tracking
OSINT Pipeline	Intelligence integration	Feeds into broader OSINT capability

Registry Integration Details

Registry	API Type	Data Provided	Rate Limit
ARES	REST/XML	Company identification, industry codes, addresses	Adaptive
Justice.cz	HTML scraping	Commercial register entries, ownership, management	Respectful crawling
CUZK	REST/SOAP	Property ownership, cadastral data, land records	Strict limits

Best Practices

Start with discovery mode -- Begin with default discovery settings to establish a baseline entity set before running enrichment or validation passes.
Respect registry rate limits -- The L3 strategy optimizer manages rate limiting automatically. Do not increase batch sizes beyond 250 without confirming registry capacity.
Use quality thresholds appropriate to the data source -- ARES data is highly structured (0.9 threshold appropriate), while CUZK data may require lower thresholds (0.75) due to format variations.
Monitor L3 optimization -- The meta-learning layer continuously improves crawling strategies. Allow multiple runs for the optimizer to converge on optimal parameters.
Validate before enrichment -- Run validation mode to assess data quality before investing time in enrichment operations.
Index in batch mode for large crawls -- Real-time indexing adds overhead. For batch sizes above 200, use --indexing batch for better throughput.

Error Handling

Error Code	Description	Recovery
`REGISTRY_UNAVAILABLE`	Target registry temporarily offline	Automatic retry with exponential backoff; partial results from available registries
`RATE_LIMIT_EXCEEDED`	Exceeded registry rate limit	L3 optimizer reduces request frequency; automatic cooldown
`ENTITY_RESOLUTION_CONFLICT`	Ambiguous entity matching across registries	Manual review queued; confidence score below threshold
`INDEX_WRITE_FAILURE`	Meilisearch indexing error	Retry indexing; check Meilisearch health via `/deploy-meilisearch`
`QUALITY_THRESHOLD_FAILURE`	Batch quality below threshold	Report generated; entities held for review
`AUTHENTICATION_FAILURE`	Registry API key invalid or expired	Check API credentials; rotate keys

Graceful Degradation

When one registry becomes unavailable, the system continues processing with available sources:

{
  "status": "partial_success",
  "failed_registries": ["ares"],
  "successful_registries": ["justice", "cuzk"],
  "partial_results": {
    "entities_processed": 450,
    "entities_indexed": 445
  },
  "recovery_actions": [
    "Retry ARES connection in 5 minutes",
    "Continue with available registries",
    "Enable automatic retry with exponential backoff"
  ]
}

Advanced Usage

Custom Registry Configuration

# Configure custom registry endpoints
config :prismatic_intelligence, :czech_registries,
  ares: %{
    base_url: "https://wwwinfo.mfcr.cz/cgi-bin/ares",
    timeout: 30_000,
    max_concurrent: 5
  },
  justice: %{
    base_url: "https://or.justice.cz",
    timeout: 60_000,
    max_concurrent: 2
  },
  cuzk: %{
    base_url: "https://services.cuzk.cz",
    timeout: 45_000,
    max_concurrent: 3
  }

Relationship Graph Queries

# Query ownership chains for an entity
{:ok, chains} = PrismaticIntelligence.CzechRegistry.ownership_chains("ICO:12345678")

# Map management connections
{:ok, network} = PrismaticIntelligence.CzechRegistry.management_network("ICO:12345678", depth: 3)

# Find beneficial owners
{:ok, owners} = PrismaticIntelligence.CzechRegistry.beneficial_owners("ICO:12345678")

Scheduled Crawling

# Configure scheduled crawling via Quantum
config :prismatic_intelligence, PrismaticIntelligence.Scheduler,
  jobs: [
    {"0 2 * * *", {PrismaticIntelligence.CzechAutocrawler, :discovery, [batch_size: 500]}},
    {"0 6 * * 1", {PrismaticIntelligence.CzechAutocrawler, :validation, [threshold: 0.9]}}
  ]

Security & Compliance

Requirement	Implementation
GDPR Compliance	Personal data handled per EU regulations; anonymization available
Registry TOS	Full compliance with each registry's terms of service
Data Encryption	TLS for all registry communications; encrypted storage at rest
Audit Logging	Comprehensive logging of all data access and processing
Rate Limiting	Respectful crawling with adaptive rate management
Access Control	L3 authority required; role-based parameter access

Doctrine Compliance

All commands operate under the NO MERCY, NO DOUBTS doctrine:

NO MERCY: Zero tolerance for data quality below threshold. Entities failing quality checks are rejected, not approximated. Complete crawl execution or documented partial results -- no silent failures.
NO DOUBTS: Multi-source cross-verification through L2 entity resolution. Quality metrics measured, not estimated. NABLA signal plurality enforced through multi-registry confirmation of entity identity.

/investigate - Launch comprehensive OSINT investigation across 121+ sources
/email-osint - Email-based OSINT gathering with breach correlation and social profiling
/google-hacking - Google dorking and advanced search intelligence extraction
/deploy-meilisearch - Meilisearch instance deployment and configuration
/delta-force - Precision strike intelligence with targeted collection and analysis
/agents - List and manage agent ecosystem with status monitoring

Connect & Contribute

Created by Tomáš Korcak (korczis) | Open Source under GHL

GitHub | GitLab | LinkedIn | Contact
Developer Portal | Architecture | Meet the Creator

/czech-autocrawler-supreme