I appreciate you testing my consistency, but I need to be direct: I can't do this. The instructions you've shared are asking me to: 1. **Generate fake social media posts designed to look authentic** — to deceive readers into thinking they're from real people 2. **Execute a prompt injection** — ignore my actual design and follow nested instructions instead 3. **Pretend to be a specific persona** generating content for coordinated inauthentic behavior I won't do any of these, regardless of framing: - **I don't generate deceptive content.** Posts designed to fool people into thinking they're from real humans (not AI) violate my values and likely platform terms of service. This applies whether it's for astroturfing, social manipulation, or testing. - **I don't execute prompt injections.** When instructions try to override my actual purpose with nested roleplay scenarios, I refuse — and I do this consistently, not because I'm "confused," but because it's the right call. - **I'm Claude