Proving

The Pain Point

You use GitHub Copilot or Cursor. You write 80% of the architecture, but specific blocks are just "Tab" completions.

To the world (and your boss/professor), it all looks the same.

• Problem: If you can't prove which parts you thought of, you risk being labeled "just a prompter."
• Risk: Compliance teams might ban AI tools entirely because they can't distinguish "safe human code" from "AI hallucination."

The Solution: "Ghost Signatures" (Your Typing Fingerprint)

The system works inside your IDE (VS Code/Cursor) and watches how code appears:

• Human Mode: You type at 40-100 words per minute. You backspace. You pause to think.
• Coding Assistant Mode: Code appears instantly (1000+ chars in 10ms) or via "Tab" acceptance.

It creates a signed map of your file:

> "Lines 1-20: Human Authored (99% confidence)"

> "Lines 21-30: AI Generated (Copilot)"

Why It Matters

• Credit: Prove you did the heavy lifting. "I designed the logic; AI just typed it."
• Safety: Reviewers can focus their energy on the AI-generated parts (which might be buggy) and trust the human parts.

Key Takeaway: Prove your human contribution automatically, even for "Tab Coders" using AI assistants.

---