Add Mistral provider support

[stamsam]

Summary

Adds Mistral as a new provider in CodexBar with a hybrid fetch strategy:

• web mode uses the Mistral AI Studio billing endpoint when a browser session is available
• api mode falls back to MISTRAL_API_KEY for model discovery and API-only status
• auto prefers web billing and falls back gracefully to the public API

What’s Included

• New Mistral provider implementation and descriptor registration
• Settings UI for usage source and optional API keys
• Automatic browser-session import for AI Studio usage when available
• API fallback when billing data is unavailable
• Mistral icon/branding
• CLI and widget wiring
• Docs updates
• Provider tests

Notes

Mistral’s documented public API does not currently expose the billing/usage totals needed for full quota cards, so real usage totals come from the logged-in AI Studio web session. The provider keeps API-key support for fallback and CLI use.

Verification

• Built CodexBarCLI
• Built CodexBar
• Repackaged and tested the app UI
• Confirmed API fallback mode
• Confirmed web billing mode can connect and return live usage data

Screenshots

Live usage in the menu card and the Mistral settings flow:

[stamsam]

** Tested both API fallback and AI Studio web usage flow locally.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 67a2194b52

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".