Docyrus E2E Browser Testing

The loop for proving a Docyrus-backed app actually works after you change it: sign in headlessly → drive the real UI → read the runtime problems the app collected. It ties together three tools:

1. docyrus auth sso-session — mint a short-lived token so the app logs in with no credential typing.
2. The docyrus browser tools — navigate, snapshot, click, fill, screenshot. Full reference: docyrus-browser-cli skill. This skill only covers the e2e-specific use.
3. @docyrus/devtools — an in-page diagnostics layer most Docyrus pi-built apps already ship. Read its collected errors/issues/console with docyrus browser devtools.

The e2e loop

1. Confirm CLI auth, env, tenant     docyrus auth who / env which / auth tenants use
2. Mint an SSO token                 docyrus auth sso-session --clientId <appClientId>
3. Navigate WITH the token           docyrus browser nav "<appUrl>?sso_token=<token>"  → app auto-signs-in
4. Drive the changed flow            docyrus browser snapshot / click / fill / wait
5. Pull the problems                 docyrus browser devtools issues|errors|console --level error
6. Triage → fix → repeat