메뉴
This skill should be used when the user asks to "create a lab module", "write a workshop module", "build a Showroom lab", "convert docs to a lab", "write a hands-on exercise", "create an AsciiDoc module", or "turn this documentation into a lab exercise".
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
This skill should be used when the user asks to "create a catalog", "build a common.yaml", "add a new RHDP lab", "set up a new catalog item", "create an AgnosticV catalog", "build a dev.yaml", "add a catalog entry", or "create a new lab catalog for RHDP".
This skill should be used when the user asks to "validate my catalog", "check my AgnosticV config", "review common.yaml", "validate my RHDP catalog", "check for errors in my catalog", "lint my AgnosticV files", or "verify my catalog before submitting".
This skill should be used when the user asks to "turn this lab into a blog post", "write a blog post from this module", "generate a blog from my Showroom content", "convert my workshop to a blog", "write a Red Hat Developer blog post", or "create a blog post from my demo".
This skill should be used when the user asks to "create a demo module", "write a Know/Show demo", "build a presenter demo", "create a Showroom demo", "write a facilitator guide", "build a demo script", or "create a presenter-led demo for RHDP".
This skill should be used when the user asks to "write solve and validate playbooks", "add E2E testing to my lab", "generate solve.yml", "generate validate.yml", "add solve and validate buttons", "set up runtime automation", "add ZT grading to my lab", "set up load testing for my lab", "add Demolition grading", or "configure run_e2e_load_test".
Infrastructure Health Checker — use this when you need to validate that a workload or service DEPLOYED SUCCESSFULLY (pods running, routes accessible, dependencies ready). This is NOT for student E2E lab testing. For student exercise validation (solve/validate buttons), use /ftl:rhdp-lab-validator instead. Triggers on: 'create a validation role', 'build health checks', 'add deployment validation', 'create post-deployment checks', 'validate my deployment', 'check if my workload is healthy'.
| name | showroom:create-lab |
| description | This skill should be used when the user asks to "create a lab module", "write a workshop module", "build a Showroom lab", "convert docs to a lab", "write a hands-on exercise", "create an AsciiDoc module", or "turn this documentation into a lab exercise". |
Orchestrates agents to create Red Hat Showroom workshop or demo content. Two modes: interactive (human) and headless (Publishing House via ph_payload).
This skill is an orchestrator. Generation and review are delegated to agents:
showroom:file-generator (Sonnet) — generates one AsciiDoc file, runs in parallel for new labsshowroom:module-reviewer (Sonnet) — quality check on generated filesThe orchestrator handles: argument parsing, repo detection, planning, agent spawning, nav merge, delivery.
See @showroom/docs/SKILL-COMMON-RULES.md for AsciiDoc rules, version pinning, image conventions, and navigation format.
If ph_payload is present in the invocation, skip ALL interactive questions and run headless. Return structured JSON.
ph_payload:
target_dir: content/modules/ROOT/pages/
mode: new | continue
previous_module: 03-module-01-pipelines.adoc # continue mode only
spec:
lab_name: OpenShift Pipelines Workshop
audience: intermediate
learning_objectives: [Deploy a pipeline, Configure triggers, Monitor builds]
business_scenario: ACME Corp needs to modernize their CI/CD pipeline
duration: 90min
module_outline: |
Module 1: Pipeline setup (~30 min)
Module 2: Triggers (~30 min)
Module 3: Monitoring (~30 min)
env:
ocp_version: "4.18"
attributes: {user: user1, password: openshift}
Headless return (JSON only, no prose):
{
"files_created": ["index.adoc", "01-overview.adoc", "02-details.adoc", "03-module-01-pipelines.adoc"],
"nav_updated": true,
"quality": {"critical": 0, "high": 0, "warnings": 2},
"warnings": ["Module has only 1 exercise, recommend adding more"]
}
/create-lab # interactive, asks all questions
/create-lab <directory> # specify target directory
/create-lab <directory> --new # new lab, skip mode question
/create-lab <directory> --continue <module> # continue from module
Validate directory exists. If empty: tell user to clone nookbag template first:
git clone https://github.com/rhpds/showroom_template_nookbag <repo-name>
Never suggest showroom-nookbag (hyphens) — always showroom_template_nookbag (underscores).
If not set by arguments:
Are you:
1. Starting a NEW lab (creates index + overview + details + first module)
2. Adding a module to an EXISTING lab
Detect content type from existing files (no question):
=== Verify sections or numbered exercises → WorkshopFor NEW lab — ask as one grouped form:
Let's plan your workshop. Answer all at once or skip any you're unsure about:
Lab name:
Target audience (e.g. Developers, SREs, Platform Engineers):
Experience level (Beginner / Intermediate / Advanced):
Main learning goal (what can learners DO when finished?):
Learning outcomes (list 3-4 skills they'll gain):
Business scenario (company name + challenge, e.g. "ACME Corp needs to modernize CI/CD"):
Duration (e.g. 30min, 1hr, 2hr):
Module breakdown (how many modules, what each covers):
OpenShift version (or leave blank for {ocp_version} placeholder):
Cluster type (multinode / SNO):
Access type (admin / multi-user htpasswd / keycloak):
UserInfo variables (paste from demo.redhat.com → My services → Details → Advanced settings):
Leave blank to use placeholder attributes {user}, {password}, {openshift_console_url}
Reference materials (paste URLs, file paths, or text — leave blank to use templates):
Writing style (optional — skip for standard Red Hat style):
Describe: "conversational, short sentences, active voice, no jargon"
OR paste 1-3 paragraphs of your own writing as an example
OR give a file path: ~/my-showroom/content/modules/ROOT/pages/03-my-module.adoc
Saved profile? Point to: ~/.claude/context/my-writing-style.md
User fills what they know, skips the rest. Orchestrator confirms the plan in one summary block:
📋 Plan confirmed:
Lab: [name] | [audience] | [duration]
Modules: [breakdown]
Environment: OCP [version] | [access type]
References: [N URLs / no references]
Generating files... (running in parallel)
For CONTINUE mode — minimal questions:
Which module are you adding? (title + brief description)
Reference materials for this module? (URLs, files, or blank)
Include troubleshooting section? [Y/n]
Ask Q0–Q3 from @showroom/skills/create-lab/references/showroom-scaffold.md:
Create/update site.yml and ui-config.yml in repo root.
Build FULL_SPEC JSON from the planning form answers.
NEW lab — spawn all files simultaneously:
Task tool:
subagent_type: showroom:file-generator
prompt: |
TARGET_FILE: <repo_path>/content/modules/ROOT/pages/index.adoc
FILE_TYPE: index
FULL_SPEC: <FULL_SPEC JSON>
LAB_TYPE: <ocp|rhel|vm|ai>
CONTENT_TYPE: <workshop|demo>
REPO_PATH: <absolute repo path>
Task tool:
subagent_type: showroom:file-generator
prompt: |
TARGET_FILE: <repo_path>/content/modules/ROOT/pages/01-overview.adoc
FILE_TYPE: overview
FULL_SPEC: <FULL_SPEC JSON>
...
Task tool:
subagent_type: showroom:file-generator
prompt: |
TARGET_FILE: <repo_path>/content/modules/ROOT/pages/02-details.adoc
FILE_TYPE: details
FULL_SPEC: <FULL_SPEC JSON>
...
Task tool:
subagent_type: showroom:file-generator
prompt: |
TARGET_FILE: <repo_path>/content/modules/ROOT/pages/03-module-01-<slug>.adoc
FILE_TYPE: module
FULL_SPEC: <FULL_SPEC JSON>
...
All four run concurrently. Wait for all to return JSON.
CONTINUE mode — single agent, sequential:
Task tool:
subagent_type: showroom:file-generator
prompt: |
TARGET_FILE: <repo_path>/content/modules/ROOT/pages/<next-module>.adoc
FILE_TYPE: module
FULL_SPEC: <FULL_SPEC JSON>
PREVIOUS_MODULE: <path to previous .adoc>
REPO_PATH: <absolute repo path>
Spawn showroom:module-reviewer on each generated file:
Task tool:
subagent_type: showroom:module-reviewer
prompt: |
MODULE_FILE: <path to generated .adoc>
CONTENT_TYPE: <workshop|demo>
LAB_TYPE: <ocp|rhel|vm|ai>
SHARED_CONTEXT: {"module_order": [...], "defined_attributes": {...}, "first_use_map": {}, "lab_type": "ocp", "content_type": "workshop"}
REPO_PATH: <absolute repo path>
is_first_module: <true if overview>
is_conclusion: false
Collect findings. If Critical or High issues found, fix inline before delivering.
Merge nav_entry values from each file-generator JSON output.
Read existing content/modules/ROOT/nav.adoc. Insert new entries in correct order. Write back.
Confirm: ✅ nav.adoc updated with [N] new entries.
Human mode:
✅ Files created:
index.adoc (247 words)
01-overview.adoc (312 words)
02-details.adoc (198 words)
03-module-01-pipeline-setup.adoc (1,340 words, 2 exercises)
✅ nav.adoc updated
Quality: 0 Critical, 0 High, 2 Warnings
- W1: Module has only 1 exercise — consider adding a second
- W2: [source,bash] found — run /verify-content E.3a fix when ready
Next: Add more modules with /create-lab --continue, or verify the full lab with /verify-content
Headless mode: return JSON from Phase 2's ph_payload schema.
When all modules are complete and user asks for the conclusion:
Task tool:
subagent_type: showroom:file-generator
prompt: |
FILE_TYPE: conclusion
FULL_SPEC: <FULL_SPEC JSON with all module summaries>
...
See @showroom/skills/create-lab/references/conclusion-template.md for conclusion structure.
/showroom:verify-content — full quality review after content is created/showroom:create-demo — presenter-led demo content/ftl:rhdp-lab-validator — write E2E solve/validate automation