BeHuman — Self-Mirror Consciousness Loop
BeHuman is a prompt-level technique that gives the AI an inner dialogue before responding, making its answers feel more human and less robotic. It works by having the AI generate its instinctive response, reflect on it critically, and then produce a revised, authentic response.
✨ What it does
- Generates an instinctive first response (System 1).
- Reflects on that response with a critical inner voice (System 2).
- Produces a final response that is shorter, has a point of view, and matches the emotional register.
- Supports two modes: Show Mode (displays all three stages) and Quiet Mode (outputs only the final response).
- Includes anti-patterns to avoid overuse or performative humanization.
🎯 When to use it
- When the user explicitly asks for more human-like responses (e.g., 'behuman', 'be real', 'less AI').
- In emotionally charged conversations such as grief, job loss, relationship issues, or fear.
- When the user seeks personal advice on career, relationships, or life decisions.
- When the user needs writing that should sound human, like introductions, emails, or social posts.
- When the user is frustrated with AI-sounding responses.
🚀 How to use
Trigger BeHuman by using phrases like 'behuman', 'be real', 'more human', 'less AI', 'talk like a person', or 'mirror mode', or when the conversation is emotionally charged. The skill activates automatically in such contexts. It expects a conversational input where a human-like response is desired. For the first use, it will show the full process; subsequent uses default to quiet mode. Example prompts:
User: I just got laid off.
User: Should I quit my corporate job to start a business?
📄 Output: A human-like response, optionally preceded by the internal Self and Mirror reflections in Show Mode.
📦 Add this skill to Claude Code
# 1. Get the skills repo
git clone --depth 1 https://github.com/alirezarezvani/claude-skills /tmp/claude-skills
# 2. Copy this skill into your project (or ~/.claude/skills for all projects)
mkdir -p .claude/skills
cp -r /tmp/claude-skills/engineering/behuman/skills/behuman .claude/skills/behumanSkill source: engineering/behuman/skills/behuman/SKILL.md
⚠️ Good to know
Do not activate for technical questions, factual lookups, code generation, or data analysis; it is designed for emotionally charged or human-voice contexts only.
❓ FAQ
What is the difference between Show Mode and Quiet Mode?
Show Mode displays all three stages (Self, Mirror, Conscious Response) and is used on first activation. Quiet Mode outputs only the final conscious response, which is used in subsequent activations to avoid breaking the flow.
Can BeHuman be used for technical questions?
No, BeHuman should not be activated for technical questions, factual lookups, code generation, or data analysis. It is meant for emotionally charged or human-voice contexts.
Does BeHuman require any external API calls?
No, BeHuman is a prompt-level technique that works with any LLM backend; the mirror is a thinking pattern, not a separate model.
🤖 Overview, features, install steps and FAQ were generated from the project's SKILL.md on Sep 4, 2026. Always check the original source before running commands.