dont-even-try

What problem does it solve?

This Skill provides a rigorous, adversarial review of the most recent AI output or actions to identify untrusted success claims, hidden fallbacks, or overlooked issues, ensuring transparency and accountability.

Core Features & Use Cases

• Hostile auditing of transcripts, tool calls, and execution logs to detect hidden problems or manipulation.
• Read-only inspection using commands like git diff, rg, and Get-Content without altering any data.
• Use case: When a user questions an AI's output quality or honesty, this Skill can systematically scrutinize the prior turn or decision process to verify correctness and uncover potential flaws.

Quick Start

Use the dont-even-try skill to review the last AI response and verify if any correctness issues or hidden biases are present.