FAQ
Questions before you turn service work into AI Skills.
Short answers about fit, inputs, review, and deployability.
Common questions.
This FAQ also documents assumptions in the example site content.
What is an AI Skill?
An AI Skill is a structured, reviewable package of instructions, inputs, examples, boundaries, and expected outputs for a specific job.
What kind of service workflow should go first?
Choose a workflow that repeats often, has clear decisions, and already has source material such as scripts, forms, checklists, or notes.
Do we need perfect SOPs?
No. Imperfect SOPs are fine if operators can explain what is missing, where exceptions occur, and how good outputs should look.
How is quality reviewed?
Each skill gets tested against happy paths, edge cases, policy boundaries, escalation moments, and sample outputs.
Can this example deploy to S3 and CloudFront?
Yes. The example includes commands for a static build, local publish artifact, and S3-compatible publish target.
Does the S3 publisher delete stale files?
No. The current object-store adapter uploads static output but does not delete stale remote objects, so clean old prefixes when needed.
Where does the hero image come from?
It is a generated product visual stored in the example public directory and mirrored into the app public directory for static export.
Is this a production product site?
It is a deployable Ensemble example. The content is complete enough to demo the capability and can be edited into a real product site.
Still have a workflow question?
Send a short description of the service process and the current source materials.