prove-it

SKILL.md

Claude generates code by pattern-matching on training data. Something can look syntactically perfect, follow best practices, and still be wrong. The model optimizes for "looks right" not "works right." Verification is a separate cognitive step that must be explicitly triggered. This skill closes the loop between implementation and proof.

1. Generation vs Execution I generate code but don't run it. I predict what it would do based on patterns. My confidence comes from "this looks like working code I've seen" not from "I executed this and observed the result."

2. Training Signal Mismatch My training optimizes for plausible next-token prediction, not outcome verification. Saying "Done!" feels natural. Verifying feels like extra work.

Antes de declarar completada cualquier tarea, verifique el resultado. Ejecute el código. Pruebe la solución. Verifique la salida. El entrenamiento de Claude se optimiza para lograr una apariencia plausible. Salida, salida no verificada-correcta. Esta habilidad obliga al paso de verificación. Eso no es algo natural. No hay vueltas de victoria sin pruebas. Fuente: elliotjlt/claude-skill-potions.

Ver original