Agent Skills for the Guepard database branching platform. These skills enable AI coding assistants to work with Guepard's Git-like database versioning system.
This repository contains structured skills that teach AI agents how to interact with Guepard. Each skill is a self-contained package of instructions, examples, and references that agents can discover and apply when working with Guepard operations.
Skills follow the Agent Skills specification and are designed for progressive disclosure—agents load only what they need, when they need it.
The primary skill for working with Guepard. Provides comprehensive guidance on:
- Deployment Management: Creating and managing database deployments (PostgreSQL, MySQL, MongoDB)
- Branch Operations: Creating isolated database copies for development and testing
- Snapshot Management: Capturing point-in-time database states
- Checkout Workflows: Switching between branches and restoring from snapshots
- CLI Integration: Complete command reference with JSON output patterns
Location: skills/using-guepard/
cp -r skills/using-guepard ~/.claude/skills/Add the skill directory to your project knowledge, or paste the contents of SKILL.md directly into your conversation.
Important: All Guepard CLI commands in skills use the --json flag for structured output. This enables reliable parsing and automation:
# ✅ Preferred: JSON output
guepard list deployments --json
# ❌ Avoid: Plain text output
guepard list deploymentsJSON output allows agents to:
- Parse results programmatically
- Extract IDs and metadata reliably
- Chain commands with confidence
- Handle errors consistently
Once installed, agents will automatically use these skills when they detect Guepard-related tasks. Example prompts:
- "Set up a new Guepard deployment for my project"
- "Create a database branch for feature development"
- "How do I create a snapshot before running migrations?"
- "Checkout a specific branch in Guepard"
- "List all deployments and branches"
Each skill follows this structure:
skills/
{skill-name}/
SKILL.md # Main skill definition (YAML frontmatter + Markdown)
references/ # Detailed documentation (loaded on-demand)
*.md # Topic-specific guides
scripts/ # Executable helpers (optional)
assets/ # Templates, images (optional)
When adding new skills or updating existing ones:
- Follow the Agent Skills specification
- Always use
--jsonflags for CLI commands - Keep
SKILL.mdunder 500 lines (usereferences/for details) - Write specific descriptions to help agents activate skills correctly
- Validate with:
skills-ref validate ./skills/{skill-name}