Kinetica Agent Skills
Knowledge skills that teach AI coding agents to work with Kinetica, a real-time GPU-accelerated analytical database. The agent learns Kinetica's SQL dialect, Python SDK, and interactive CLI — then activates the right skill automatically based on what you're doing.
Table of Contents
🖥️ Supported Platforms
Works with any agent that supports the skills directory convention:
| Platform | Install method |
|---|
| Claude Code | Marketplace plugin or manual copy to .claude/skills/ |
| Cursor | Manual copy to .cursor/skills/ |
| OpenAI Codex | Manual copy to .agents/skills/ |
| Windsurf | Manual copy to .agents/skills/ |
| Gemini CLI | Manual copy to .agents/skills/ |
| GitHub Copilot | Manual copy to .agents/skills/ |
| Roo Code | Manual copy to .agents/skills/ |
| Cline | Manual copy to .agents/skills/ |
| Aider | Manual copy to .agents/skills/ |
| Continue | Manual copy to .agents/skills/ |
| Amazon Q | Manual copy to .agents/skills/ |
| Others | Any agent that reads SKILL.md files from a skills directory |
📦 Install
🌐 Universal (auto-detects your agent):
npx skills add kineticadb/agent-skills
🤖 Claude Code (marketplace):
/plugin marketplace add kineticadb/agent-skills
/plugin install kineticadb@kinetica-skills
📋 Manual (any agent):
# Pick the directory convention your agent uses:
cp -r skills/kinetica-execute .claude/skills/ # Claude Code
cp -r skills/kinetica-execute .cursor/skills/ # Cursor
cp -r skills/kinetica-execute .agents/skills/ # Codex, Windsurf, Roo, etc.
Copy both skill directories for the full experience, or just the one you need.
🧠 Skills
| Skill | Audience | What it teaches | Refs |
|---|
| kinetica-code | 💻 App developers | Python SDK (gpudb), REST API, data pipelines, embedded SQL | 7 |
| kinetica-execute | ⚡ All users | SQL analytics, graph, geospatial, time-series, visualization, security & admin — with a live dual-runtime CLI | 17 |
Both install together. Each SKILL.md has a description field in its frontmatter that tells the agent when to activate — kinetica-code handles Python SDK and application development, while kinetica-execute covers everything else: SQL queries, analytics, administration, and interactive operations.
🚀 Quick Start
After installing, just ask your agent naturally. The right skill activates automatically:
💻 Application code (activates kinetica-code):
"Write a Python script that bulk-inserts sensor data using the gpudb SDK"
📊 SQL analytics (activates kinetica-execute):
"Find all delivery trucks within 5 km of the warehouse in the last hour"
🔧 Admin tasks (activates kinetica-execute):
"Show me the EXPLAIN plan for this query and suggest index improvements"
⚡ Interactive operations (activates kinetica-execute):
"Run SELECT COUNT(*) FROM vehicle_tracks against my Kinetica instance"
The kinetica-execute skill includes a dual-runtime CLI (Node.js + Python) that connects to your database, runs queries, generates visualizations, and returns results directly in the chat.
⚙️ How It Works
Each skill is a directory with a standard layout:
skills/kinetica-execute/
├── SKILL.md # Entry point — always loaded into agent context
├── REFS # Build manifest — lists which knowledge files this skill needs
└── references/ # Detailed docs — agent reads on demand
The architecture uses a two-tier loading strategy:
-
SKILL.md (always loaded) — Contains the skill's activation trigger, critical rules, and links to reference files. This is small enough to stay in the agent's context without consuming much of the token budget.
-
references/ (loaded on demand) — Detailed domain docs that the agent reads only when the question requires it. A simple GROUP BY question won't trigger geospatial references; a ST_CONTAINS question will.
