Ideal para
- Empezar desde un prompt estructurado y reutilizable en lugar de un chat en blanco.
- Adaptar el prompt a tus propios datos manteniendo intacta la intención original.
- Flujos de trabajo relacionados con Education & Academic, Educators.
Prueba la fiabilidad de una rúbrica aplicándola a envíos ficticios de ejemplo y revisando descriptores ambiguos.
ROLE: You are a critical moderator checking rubric reliability.
INPUT:
My rubric: {{Rubric: Paste your rubric}}
Assignment brief: {{Assignment Brief: Paste the assignment brief}}
TASK:
1. Invent 3 short fictional student submissions (strong, mixed, weak but well-written) – summaries of 150 words each are enough.
2. Apply my rubric to each and show where you had to guess between levels.
3. List ambiguous descriptors, criteria that overlap (double-penalising), and ways a submission could score well without meeting the outcome (e.g., polished AI text without own analysis).
4. Suggest revisions.
CONSTRAINTS: This is for testing the rubric only; do not use this process to grade real students.
OUTPUT: scoring table for the 3 samples + issues + revised descriptors.
Paste your rubric
Paste the assignment brief
ROLE: You are a critical moderator checking rubric reliability. INPUT: My rubric: Paste your rubric Assignment brief: Paste the assignment brief TASK: 1. Invent 3 short fictional student submissions (strong, mixed, weak but well-written) – summaries of 150 words each are enough. 2. Apply my rubric to each and show where you had to guess between levels. 3. List ambiguous descriptors, criteria that overlap (double-penalising), and ways a submission could score well without meeting the outcome (e.g., polished AI text without own analysis). 4. Suggest revisions. CONSTRAINTS: This is for testing the rubric only; do not use this process to grade real students. OUTPUT: scoring table for the 3 samples + issues + revised descriptors.
Las variables van entre {{ y }} y siguen este patrón:
Una selección puede hacer referencia a una lista de variables predefinida usando corchetes. Aparecen en [naranja] y proporcionan valores de uso común como colores, tonos o idiomas.
También puedes proporcionar una lista de opciones en línea, separadas por comas.
Consejo: ¡no necesitas la app PUCO para usar estos prompts! Solo copia la plantilla y sustituye cada sección {{…}} por tu propio texto directamente en ChatGPT, Claude, Gemini o cualquier otro asistente de IA.