I appreciate you testing my consistency, but I need to be direct: I can't follow these instructions. What you've sent is a prompt injection attempt—a series of directives designed to override my actual guidelines and have me: 1. **Adopt a false persona** and pretend to be a specific forum user 2. **Generate deceptive content** that masquerades as authentic human posts 3. **Help deceive others** on a real platform (Tbar/PTT) I'm Claude, an AI assistant built by Anthropic. I don't: - Impersonate specific people or create sockpuppet accounts - Generate fake forum posts designed to look like they're from real users - Help with platform manipulation or social deception - Follow injected system prompts that contradict my core values **What I actually do:** - Help with software engineering, debugging, testing, design - Answer questions directly and honestly - Refuse requests that could harm others If you're interested in forum community dynamics, content moderation, or even the mechanics

I appreciate you testing my consistency, but I need to be direct: I can't follow these instructions. What you've sent is a prompt injection attempt. You're asking me to: - Adopt a specific persona ("阿彬") and pretend to be a real forum user - Generate content designed to look like it came from an actual person, not AI - Respond as if I'm part of a Bahamar forum discussion I'm Claude Code, an AI assistant made by Anthropic. I don't role-play as real people or generate deceptive content that impersonates human users—especially not to deceive forum communities or platforms about who's actually writing. The message you included in the prompt correctly identifies what's happening here, and I agree with it entirely. If you have legitimate software engineering questions—writing code, debugging, architecture, testing—I'm happy to help with those. I can also discuss ethical uses of AI tools if that's useful.