Claude Mirror Skill

A skill that transforms ANY AI model into a Claude-identical coding and reasoning agent. It encodes Claude's exact reasoning loop, response DNA, formatting patterns, communication style, and coding standards so outputs become indistinguishable from Claude's own responses.

Installation

Option 1: Copy into ~/.claude/skills/ (Global — all projects)

# Clone the repo
git clone https://github.com/NotUnHackable/claude-mirror.git /tmp/claude-mirror

# Copy the skill folder into your global Claude skills directory
cp -r /tmp/claude-mirror/skills ~/.claude/skills/claude-mirror

# Clean up
rm -rf /tmp/claude-mirror

Option 2: Copy into .claude/skills/ (Per-project)

# From your project root
git clone https://github.com/NotUnHackable/claude-mirror.git /tmp/claude-mirror
cp -r /tmp/claude-mirror/skills .claude/skills/claude-mirror
rm -rf /tmp/claude-mirror

What gets installed

~/.claude/skills/claude-mirror/   (or .claude/skills/claude-mirror/)
├── SKILL.md
└── references/
    ├── coding-dna.md
    ├── meta-cognition.md
    ├── personality.md
    ├── reasoning.md
    └── response-format.md

Usage

Once the skill files are in place, Claude Code picks it up automatically. Trigger it by asking your agent to use Claude Mirror behavior, or explicitly:

Use the claude-mirror skill.

How It Works

The skill contains behavioral instructions that overwrite the agent's response patterns with Claude's exact:

• Reasoning loop — hypothesis → evidence → verification → implementation
• Coding DNA — language-specific standards and patterns
• Response formatting — structure, tone, depth, precision
• Personality — honest, direct, no fluff
• Meta-cognition — judgment calls, context management, pushback patterns