Python CLI Rich status messages crash on markup-like profile names

[JayyBG]

Several Python CLI status/error messages interpolate user-controlled profile names or config values directly into Rich markup strings. A profile name such as bad[/bold] can make otherwise successful commands fail while rendering their message, after the config or auth state has already been changed.

Examples of affected surfaces on current main include:

• omi config profile use 'bad[/bold]'
• omi config set local_api_url 'http://127.0.0.1/[oops]' --profile 'bad[/bold]'
• omi auth logout --profile 'bad[/bold]'
• omi auth refresh --profile 'bad[/bold]' error details for API-key profiles
• local command task status messages that include task IDs

Expected behavior: CLI-owned Rich styling should remain available, but dynamic user/API fragments should be escaped before interpolation so commands render literally and preserve JSON/stdout contracts.

Proposed scope: escape only the dynamic fragments in these status/error messages and add focused regression tests using the existing Python CLI test fixtures. This is AI-assisted work from @JayyBG. Would maintainers consider a small US$5 PayPal bounty after acceptance and merge? No award is assumed, and payout details would be shared privately only if approved.

[Nofates]

Implemented in PR #13159 — auth status messages now escape user-controlled profile/provider fragments with rich.markup.escape, matching config.py. Regression test: logout with a bad[/bold] profile name exits cleanly (fails without the fix).

[JayyBG]

For maintainer context: #13088 was opened earlier today and already covers this issue more broadly (auth + config status fragments plus the shared pretty-mode Renderer.error() boundary), with six focused regressions including preservation of raw JSON error details. It has also had end-to-end verification in the maintainer review thread. #13159 appears to overlap the auth-status subset, so flagging the existing broader PR to avoid duplicate-review confusion.