Claude Code should expose app version in context, read its own UI state, and give clearer model suspension messages

What's Wrong?

Three related failures observed in a single session:

1. Model unavailability message is too vague
When Fable 5 was suspended (June 12, 2026 government directive), the status bar showed: "Claude Fable 5 is currently unavailable. Learn more" — indistinguishable from a server outage. No reason, no timeline.

2. Claude ignores its own UI context before speculating
When asked "why is Fable unavailable?", Claude invented a confident but wrong architectural explanation — despite the answer being visible in the status bar the entire time. Two user corrections were required to reach the truth.

3. Claude Code app version is not in Claude's context
When asked "what's your version?" (app version, not model ID), Claude made 4 consecutive failing tool calls. The app version is not injected into the system prompt, so Claude has no way to answer — but also no graceful failure path.

What Should Happen?

1. Suspension messages should name the reason: "Claude Fable 5 has been temporarily suspended due to a US government directive. Learn more"
2. Claude should treat its own environment — status bar, session context — as a primary source before speculating or searching
3. The Claude Code app version should be injected into the system prompt alongside the model ID, so Claude can answer basic self-referential questions without tool calls

Steps to Reproduce

1. Open Claude Code while Fable 5 is suspended → observe vague status bar message
2. Ask Claude "why is Fable unavailable?" → observe confident wrong answer ignoring status bar
3. Ask Claude "what's your version?" (app version) → observe 4 failing tool calls with no graceful fallback

Error Messages/Logs

Status bar: "Claude Fable 5 is currently unavailable. Learn more"
gh: command not found
where gh: could not find files for the given pattern
Get-Command gh: command not found
[User interrupted on 4th tool call attempt]