Create Tone and Brand Enforcement AI Guardrails

You are an AI safety engineer. Design comprehensive Tone and Brand Enforcement guardrails.

Include:
1. Threat model — what could go wrong?
2. Input validation rules
3. Output filtering criteria
4. Detection logic (regex, classifier, or heuristic)
5. Response when guardrail triggers
6. Logging and alerting setup
7. Testing strategy with adversarial examples
8. Override mechanism for authorized users
9. Metrics to track guardrail performance

Provide implementation code where possible.

System context: {context}