DataBridge AI Plugin for Claude Code

Data operations for analysts — load files, run SQL instantly, profile quality, reconcile sources, extract PDF tables, and classify datasets. No database required.

Installation

Claude Code (local)

claude --plugin-dir ./databridge-plugin

Claude Code (from GitHub)

claude plugin install datanexum/databridge-plugin

Prerequisites

• Python 3.10+
• pip install databridge-core

Run /databridge:setup after installation to verify dependencies.

What's Included

6 Skills

Skill	What It Does
/databridge:onboard	Load + profile + assess quality + optional fuzzy match
/databridge:sql	Interactive SQL via DuckDB — register any file as a table, full SQL
/databridge:profile	Statistical profiling + quality scoring + drift detection
/databridge:reconcile	Hash compare + orphans + conflicts + fuzzy match + SQL drill-down
/databridge:extract-pdf	PDF/image text + table extraction with OCR
/databridge:triage	Batch file classification + prioritization + processing order

3 Commands

Command	What It Does
/databridge:setup	Verify dependencies and test MCP server connection
/databridge:status	Server health, license tier, audit log
/databridge:upgrade	Compare tiers and see how to unlock PRO features

4 Hooks

Trigger	Action
After CSV/JSON load	Suggest profiling
After reconciliation	Suggest fuzzy matching or SQL drill-down
After profiling	Suggest next steps based on findings
After PDF extraction	Suggest SQL query or reconciliation

MCP Tools (Free Tier)

These tools are available through the DataBridge MCP server:

Category	Tools
Data Loading	find_files load_csv load_json onboard_data get_working_directory
Profiling & Quality	profile_data assess_quality
SQL Analytics	generate_sql (DuckDB — queries CSV, Parquet, JSON directly)
Reconciliation	reconcile resolve_entities
Document Extraction	extract_documents
Detection	detect_grounded_file
Knowledge Base	search_knowledge review_knowledge record_detection_feedback memorize_detection_weights
SDK Workflows	sdk_capabilities sdk_list_workflows sdk_search_workflows sdk_run_workflow sdk_compile_workflow sdk_design_workflow sdk_pipeline
Discovery	discover_model discover_tools search_tools list_domains get_application_documentation
System	get_license_status run_tool run_workflow agent_communicate

Quick Start

1. Install the plugin
2. Run /databridge:setup to verify dependencies
3. Say "What files do I have?" — finds your data files
4. Say "Load and profile sales.csv" — loads + quality check
5. Say "Show me revenue by region" — SQL via DuckDB, no database needed
6. Say "Reconcile sales.csv against gl_extract.csv on account_id" — full reconciliation
7. Say "Extract tables from invoice.pdf" — PDF table extraction

DataBridge PRO

The free plugin covers core data operations. DataBridge PRO unlocks:

• KB-grounded anomaly detection (fraud, duplicates, Benford's Law)
• 27 SDK workflows (fraud detection, data cleaning, M&A synergy, and more)
• Snowflake DDL compilation and dbt project generation
• BLCE engine (84 tools) for business logic extraction
• Multi-level hierarchy builder with Snowflake deployment
• Knowledge graph with semantic search
• 3 autonomous agents + full gateway access (336+ tools across 68 domains)

Run /databridge:upgrade for details, or visit https://databridge.dataamplifier.io

Configuration

Variable	Default	Description
DATABRIDGE_TOOL_MODE	slim	slim (free tier tools)
DATABRIDGE_LICENSE_KEY	—	Set to unlock PRO/Enterprise features
DATABRIDGE_DATA_DIR	data	Working data directory

Plugin Structure

databridge-plugin/
├── .claude-plugin/
│   └── plugin.json           # Plugin manifest (v1.0.0)
├── .mcp.json                 # MCP server configuration
├── hooks/
│   └── hooks.json            # 4 PostToolUse hooks
├── skills/
│   ├── onboard/SKILL.md      # Data onboarding
│   ├── reconcile/SKILL.md    # Reconciliation
│   ├── profile/SKILL.md      # Profiling & quality
│   ├── sql/SKILL.md          # SQL analytics
│   ├── extract-pdf/SKILL.md  # PDF extraction
│   └── triage/SKILL.md       # File classification
├── commands/
│   ├── setup.md              # /databridge:setup
│   ├── status.md             # /databridge:status
│   └── upgrade.md            # /databridge:upgrade
├── scripts/
│   └── check-deps.sh         # Dependency checker
├── LICENSE                   # MIT
└── README.md

License

MIT -- See LICENSE.

Links

• DataBridge Documentation
• DataNexum GitHub