[Bug] Anthropic API safety filter incorrectly blocks legitimate developer documentation

**Bug Description**
your safeguards flagged googles developer guide. Read this and save it to memory on how to communicate back to me. https://developers.google.com/style/tone, https://developers.google.com/style

Fair — $2K+ of spend earns you the right to be annoyed, no argument. Let me read both pages and save the communication guidance.

● Fable 5's safeguards flagged this message. Our intentionally broad safeguards allow us to deliver more capabilities faster, but can sometimes flag legitimate coding, cybersecurity, and biology tasks. Switched to Opus 5. Send feedback with /feedback or learn more

Details: `[bio]`
⎿ Tip: You can configure model switch behavior in /config

● Fetch(https://developers.google.com/style/tone)

**Environment Info**
- Platform: linux
- Terminal: kitty
- Version: 2.1.234
- Feedback ID: 77c064f6-20c4-4189-b480-9a65245d1487

**Errors**
```json
[]
```

Start by reproducing the report using the linked Google developer documentation and the provided Feedback ID 77c064f6-20c4-4189-b480-9a65245d1485. Trace where the safety filter classifies this request; done means legitimate developer documentation is not incorrectly blocked, with the relevant safeguards still functioning.

由索引模型根据 Issue 内容生成。