AI-powered assistance for building Oxy analytics projects. Includes skills for semantic layer development, workflow creation, ETL pipelines, and data app building.

Available for Claude Code (plugin) and Cursor (project rules).

# Install from marketplace
/plugin marketplace add oxy-hq/skills
/plugin install oxy-skills@oxy-hq

# Or use locally
claude --plugin-dir /path/to/oxy-template

Copy the project rules into your Oxy project:

# From this repo
./cursor/install.sh /path/to/your-oxy-project

# Or manually copy
cp cursor/rules/*.mdc /path/to/your-oxy-project/.cursor/rules/

See cursor/README.md for full setup instructions.

oxy-semantic-layer Builds semantic layer files (views and topics) for natural language analytics.

• Analyzes database schemas from .databases/
• Creates *.view.yml files with entities, dimensions, and measures
• Generates *.topic.yml files to organize views
• Validates with oxy build

oxy-workflow-builder Creates data workflows, SQL queries, and AI agents following Oxy patterns.

• Designs multi-step ETL pipelines
• Generates parameterized SQL with Jinja2
• Creates *.workflow.yml and *.agent.yml files
• Tests with oxy run

oxy-etl-builder Builds ETL pipelines using DLT (data-load-tools) for loading data into warehouses.

• Sets up API connectors (Toast, Square, Stripe, etc.)
• Creates spreadsheet/file ingestion pipelines
• Generates etl/sources/, etl/runners/, and transform files
• Supports ClickHouse, Snowflake, MotherDuck, DuckDB

oxy-app-builder Creates data apps (*.app.yml) - interactive dashboards combining tasks and visualizations.

• Builds SQL, workflow, semantic query, and agent tasks
• Renders outputs as tables, charts (line, bar, pie), and markdown
• Supports plan-first workflow with user approval
• Validates task-to-display data references

oxy-test-drafter Bootstraps and refines .test.yml eval files for Oxy agents and workflows.

• Scaffolds test files from a list of prompts
• Runs oxy test --output-json and parses multi-run JSON traces
• Drafts concise, evidence-based expected strings
• Classifies cases as stable, flaky, ambiguous, or unsupported
• Outputs a diagnostic summary with per-case classification

config-validator Autonomous validator that checks Oxy YAML files for syntax errors, schema violations, and reference issues.

.claude-plugin/
├── plugin.json          # Plugin manifest
└── marketplace.json     # Marketplace metadata

skills/
├── oxy-semantic-layer/
│   ├── SKILL.md         # Skill instructions for Claude
│   ├── README.md        # User-facing documentation
│   ├── QUICK-REFERENCE.md
│   └── *.yml            # Templates
├── oxy-workflow-builder/
│   ├── SKILL.md         # Workflow skill instructions
│   ├── README.md        # User documentation
│   ├── QUICK-REFERENCE.md
│   └── *-template.*     # SQL, workflow, agent templates
├── oxy-etl-builder/
│   ├── SKILL.md         # ETL skill instructions
│   ├── README.md        # User documentation
│   ├── QUICK-REFERENCE.md
│   ├── playbook-*.md    # Source-specific guides
│   ├── etl-style-guide.md
│   ├── warehouse-modeling.md
│   └── templates/       # Code templates
├── oxy-app-builder/
│   ├── SKILL.md         # App builder skill instructions
│   ├── README.md        # User documentation
│   ├── QUICK-REFERENCE.md
│   ├── templates/       # App templates
│   └── examples/        # Example prompts
└── oxy-test-drafter/
    ├── SKILL.md         # Test drafter skill instructions
    └── README.md        # User documentation

commands/
├── validate.md
├── build.md
├── sync.md
└── test.md

agents/
└── config-validator.md  # Autonomous agent instructions

cursor/                      # Cursor IDE support
├── rules/
│   ├── oxy-core.mdc         # Always-on core conventions
│   ├── oxy-semantic-layer.mdc
│   ├── oxy-workflow-builder.mdc
│   ├── oxy-etl-builder.mdc
│   └── oxy-app-builder.mdc
├── install.sh               # Copy rules into a project
└── README.md                # Cursor setup guide

• Claude Code CLI
• Oxy CLI (oxy --version)
• An Oxy project with config.yml

Skill files (skills/*/SKILL.md):

---
name: skill-name
description: When this skill activates (user-facing)
---

# Skill Title

Instructions for Claude on how to execute this skill...

Command files (commands/*.md):

---
name: command-name
description: What this command does
---

#!/bin/bash
# Command implementation

Agent files (agents/*.md):

---
name: agent-name
description: When this agent should activate
tools: [Read, Grep, Bash]
---

# Agent Instructions

Autonomous agent behavior...

See CLAUDE.md for guidance on developing this plugin with Claude Code.

For issues or feature requests: GitHub Issues

MIT - see LICENSE

• Oxy Documentation
• Claude Code Plugins Guide
• Plugin Development Skill