Reviewing Your Own Code

The hardest code to review is your own. You have blind spots because you wrote it. AI provides a fresh perspective without any assumptions about what you intended.

The Self-Review Prompt

Catching Your Blind Spots

Ask questions you might not think to ask yourself:

Missing Error Handling

"What happens if [API call] fails? Am I handling all error cases?"

Implicit Assumptions

"What assumptions is this code making about the input? Are they documented?"

Edge Cases

"What inputs would break this function? Empty arrays, null values, negative numbers?"

Race Conditions

"Could anything go wrong if this runs concurrently?"

The Devil's Advocate Prompt

Questions to Ask AI About Your Code

• "What would a senior developer criticize about this?"
• "If this caused a production incident, what would be the likely cause?"
• "What's the weakest part of this implementation?"
• "What would make this harder to maintain in 6 months?"

Key Takeaway

When reviewing your own code, explicitly ask AI to be critical and skeptical. You need someone to challenge your assumptions, not confirm them.