[MODEL] Fable-5 is practically useless for anything code-related

Preflight Checklist

• [x] I have searched existing issues for similar behavior reports
• [x] This report does NOT contain sensitive information (API keys, passwords, etc.)

Type of Behavior Issue

Claude refused a reasonable request

What You Asked Claude to Do

Fable-5 is practically useless for any technical task with even the slightest connection to code — it halts the running task outright and switches me to Opus 4.8, with a message saying its safety measures flagged the prompt for "cybersecurity or biology" and admitting they "may flag safe, normal content as well." They do, constantly. It trips on basically any backend work — anything touching authorization, networking, crypto libraries, parsing, input validation, sessions — i.e. normal software.

What Claude Actually Did

The problem is that Fable-5 can't do the thing Claude Code exists to do. It's a coding tool, and the flagship model bails on ordinary coding work. You've clearly over-indexed on safety to the point of breaking the product.

Expected Behavior

So what topics is this new model actually allowed to answer — "hi, how are you"? Are you building a tool for housewives? Whoever came up with this restriction isn't very bright and has clearly never written a single line of backend code in their life.

Permission Mode

Accept Edits was ON (auto-accepting changes)

Can You Reproduce This?

Yes, every time with the same prompt

Claude Model

Fable-5

Claude Code Version

2.1.170 (Claude Code)

Additional Context

I'm a Max (20x) subscriber paying $200/month.