docs(storage-fsx-windows-sla-optimizer): link the deployable demo - #116
Conversation
…the GenAI Ops Demo Library
ams-thakkar
left a comment
There was a problem hiding this comment.
Approving. Every claim in the 19 added lines is falsifiable, and I checked them against the demo rather than reading for plausibility: the link returns 200, all three scenario headings match lab/scenarios.yaml exactly, each carries an autoRevertSeconds and a Rollback path so "reversible" holds, and MISCONFIGURED, FreeStorageCapacity, Single-AZ, the self-managed AD and "about an hour" all match the demo's own README.
The one that mattered most to me — "references the skill at a stated ref of this repository and never copies it" — is true: shared/devops-agent/agent-tools/deploy-skill.sh sparse-clones this repo and zips the skill directory, and deploy-all.sh passes --ref main explicitly. No vendored copy anywhere in that repo, so the demo can't drift from what we ship here.
"With and without this skill" is recorded as a table column per scenario in the demo README, so that line is accurate too.
Repo side: mkdocs build --strict clean, the new link is absolute and consistent with the earlier absolute-link fixes in bc50051 and 6075d1e, no CHANGELOG bump needed for a README-only change per ample precedent, and zero commits behind main.
Thanks for pointing the skill at something people can actually run.
Summary
Adds a "Try it in a deployable demo" section to the
storage-fsx-windows-sla-optimizerREADME, pointing at the demo that showcases this skill in the GenAI Ops Demo Library: resilience/fsx-windows-sla-review-devops-agent.The demo deploys an Agent Space with this skill registered (through the Asset API), a Single-AZ FSx for Windows file system joined to a self-managed Active Directory, its alarms, and a Lab that injects three reversible mis-configurations (rotated AD credentials, backups disabled, alarm removed) and records what the agent concludes with and without the skill. The demo references the skill at a stated ref of this repository and never copies it; its README links back here. This PR closes the loop in the other direction.
README-only change: no skill content, evals or front matter touched. The eval evidence already present in the skill directory keeps the PR check green.
Testing
lab/scenarios.yaml).By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.