Hermes repairs Shopify schema after backing it up

Jim Stewart describes a Hermes setup with specialists for Shopify engineering, search engineering and client-meeting preparation. Each has client context and its own skills. The agency's SKAW tool surfaces information from client sites, giving the agents a concrete starting point for investigating changes rather than a generic optimisation request.

After a client's product snippets and related metrics dropped, the agents diagnosed a schema problem, proposed a repair and applied it. Stewart says a backup procedure runs before any client-site change, with those backups also pushed to GitHub daily. His account identifies the sequence as diagnosis, proposal, backup and fix, although it does not specify a separate human approval click for this particular repair.

Afterwards, SKAW showed the Google Shopping bot returning. Stewart explicitly calls this correlation and says he needs observations from another two or three sites to confirm the relationship. He also notes that the change was not reflected in Google Search Console's crawl-bot traffic. The reported outcome therefore supports a completed, recoverable schema repair and an observed crawl signal. It does not establish improved sales or prove that the repair caused the bot's return. The useful operational pattern is to preserve a restore point and identify which measurement system actually supplies the result.

Prompt
Suggested prompt (written by Agents In Use, not the author's)

Source: Blog · Jim Stewart