RHDP Skills Marketplace
AI-powered skills for Claude Code. Create workshops, demos, and catalog items faster — with built-in Red Hat standards and quality checks.
Available Skills
8 skills · 10 agents · 4 namespaces — orchestrator pattern with parallel execution
Showroom — Workshop & Demo Creation
Create labs, demos, and blog posts for Red Hat Showroom. Handles AsciiDoc, navigation, quality validation, and Red Hat style standards.
/showroom:create-lab— Hands-on workshop modules (Know/Do/Check)/showroom:create-demo— Presenter-led demos (Know/Show)/showroom:verify-content— Quality & standards validation/showroom:blog-generate— Workshop → blog post
AgnosticV — Catalog Automation
Build and validate RHDP catalog items — common.yaml, dev.yaml, description.adoc, info messages — with guided questions covering all infrastructure types.
/agnosticv:catalog-builder— Create or update catalog items/agnosticv:validator— Validate before submitting PR
Health — Infrastructure Validation
Did my workload deploy correctly? Generates Ansible roles that check pods, routes, and services are healthy — run at provisioning time, not during student exercises.
/health:deployment-validator— Create Ansible health check roles for deployed workloads
FTL — E2E Lab Testing
Can a student complete the lab? Reads your .adoc modules and writes solve.yml + validate.yml playbooks for student exercise validation. 4-agent pipeline with self-healing Playwright and screenshot evidence.
/ftl:rhdp-lab-validator— Orchestrate 4 agents: read → solve → validate → test/ftl:content-reader— AsciiDoc reader with vision analysis/ftl:solve-writer— Intent-based Playwright, self-healing on UI changes/ftl:env-connector— Live test + screenshot evidence + UI drift recovery
Developer Guide New
Every file that skills and agents read — how to add checks, update rules, change templates, and contribute new agents. Covers AgV validator, showroom prompts, and E2E templates.
- How skills and agents read rules (priority chain)
- Every prompt, template, and schema file explained
- How to add checks, agents, and diagrams
- PR checklist before contributing
Quick Install
Two commands to get started on Claude Code
Claude Code Recommended
Native plugin marketplace — automatic updates, version management.
/plugin marketplace add rhpds/rhdp-skills-marketplace
/plugin install showroom@rhdp-marketplaceCursor
One-command bash install. Works alongside your existing Cursor setup.
curl -fsSL https://raw.githubusercontent.com/rhpds/rhdp-skills-marketplace/main/install-cursor.sh | bashAlso works with VS Code using the Claude extension — same setup as Claude Code.
What's New — v2.14.0
Major release: agent orchestrator pattern, parallel execution, PH integration · Full changelog →
Skill-as-Orchestrator Pattern
All showroom skills now delegate to specialized agents running in parallel — same design as the FTL plugin. 6× faster on a 6-module lab (8 min → ~90 sec). Each agent returns dimension-scored JSON, enabling future regression detection.
5 New Showroom Agents New
scaffold-checker (Haiku), module-reviewer (Sonnet), file-generator (Sonnet), score-aggregator (Haiku), doc-writer (Sonnet). Skills and agents use the right model for each task — Haiku for reading, Sonnet for generation.
Publishing House Integration New
verify-content and create-lab support headless ph_payload mode. PH passes a JSON spec — the skill runs all agents silently and returns structured JSON. Zero changes to PH required.
Security Checks New
Two new validator checks based on real production issues: multiuser htpasswd labs with shared passwords, and VS Code workloads with no authentication. Both flag as High — blocking for lab readiness.
Personal Writing Style New
Describe your writing style, paste example paragraphs, or save a profile to ~/.claude/context/my-writing-style.md. All content creation skills apply your style and run an auto-humanizer pass — no AI writing patterns in the output.
babylon.yaml Schema Authority New
The agnosticv validator now reads $agv_path/.schemas/babylon.yaml as the authoritative source before running checks. Category enums, field types, and additionalProperties: false enforcement all derived from the real schema — no more invented rules.
Resources
Best Practices
Context management, CLAUDE.md setup, git rules, and real pitfalls from daily RHDP work.
Read before you start →Quick Reference
All install commands, skill commands, and common workflows in one place.
Open quick reference →Build Your Own Skill
Have a repetitive workflow? Turn it into a reusable Claude Code skill. Walkthrough guides for different roles.
Start building →Skills vs Agents explainer →
Agent Architecture
How skills and agents work together — orchestration diagrams, model assignments, agent communication patterns.
See the architecture →Publishing House Integration New
Sequence diagrams showing agent communication in interactive vs headless mode. ph_payload schema and JSON response format for all supported skills.
Integration guide →Personal Writing Style New
Save your writing style profile once and all content creation skills use it automatically. Full guide with examples.
Set up your style →