pah-synclisted

Populate this environment's Private Automation Hub as code — all Red Hat certified and validated collections windowed to their 3 newest versions, plus the ericcames and mlowcher61 community namespaces at their current version only. Regenerates the pinned lists from upstream, runs playbooks/sync_hub.yml, then asks the hub what actually landed. TRIGGER when: the user asks to populate, sync, fill or refresh Private Automation Hub or PAH, wants certified or validated collections in their hub, asks why a hub is empty or a collection is missing from it, wants to refresh hub/*-requirements.yml, or hits a hub sync that reports success with no content. SKIP: if the user wants to point AAP projects AT the hub via organization Galaxy credentials — that is pah-link-aap — or wants to mirror the execution environment image, which is sales-demos-ee-build.
ericcames/sales.demos · ★ 1 · Data & Documents · score 67

Install: claude install-skill ericcames/sales.demos

# pah-sync Populates the Private Automation Hub that ships with every AAP environment, from configuration in git rather than from the web UI. This skill contains **no logic**. The work is [`playbooks/sync_hub.yml`](../../../playbooks/sync_hub.yml) and [`utilities/refresh-hub-requirements.py`](../../../utilities/refresh-hub-requirements.py). **Every build already does most of this.** `config.yml` — stage 2 of `setup.yml` — applies the same remotes and repositories and *starts* a sync without waiting. What this skill adds is waiting for it and then proving it. ## What lands in the hub | Repository | Content | Versions | |---|---|---| | `rh-certified` | every Red Hat certified collection (214) | 3 newest of each | | `validated` | every Red Hat validated collection (47) | 3 newest of each | | `community` | the `ericcames` and `mlowcher61` namespaces (15) | current only | The lists are committed: `hub/certified-requirements.yml`, `hub/validated-requirements.yml`, `hub/community-requirements.yml`. **Pulp has no "keep N versions" control**, which is why those files exist. A requirements entry of a bare `namespace.name` syncs *every* published version, and some certified collections have 40+. Each entry instead carries an explicit version or a `>=` floor. `retain_repo_versions` is not a substitute — it prunes repository snapshots, not collection versions. ## This is laptop-only, and there is no job template Unlike every phase of the OpenShift Virtualization demo, this has on