Skip to main content

reasoning-receipt

Explain reasoning from first principles in plain words after resolving a conflict (merge conflicts, contradictory requirements, sources that disagree, competing debugging hypotheses, stakeholder trade-offs) or after giving a non-trivial recommendation, answer or analysis. Produces a short receipt naming the competing forces, the deciding principle, the assumptions, and what would break the conclusion.

Datos de origen

Repositorio
pipeshub-ai/sharp
Última actividad en el origen
15 de octubre de 2026 a las 14:15
Idioma detectado de SKILL.md
inglés
Estrellas
4
Forks
0

Opciones de instalación

De forma predeterminada está seleccionado el prompt que primero revisa el origen. Puedes cambiar a un comando directo o descargar una copia local.

Revisa los archivos de origen

Lee SKILL.md y los archivos complementarios que muestra SkillsMP antes de decidir si quieres instalarlo.

Mostrando SKILL.md

SKILL.md
Instrucciones de origen · Vista previa de solo lectura
name
reasoning-receipt
description
Explain reasoning from first principles in plain words after resolving a conflict (merge conflicts, contradictory requirements, sources that disagree, competing debugging hypotheses, stakeholder trade-offs) or after giving a non-trivial recommendation, answer or analysis. Produces a short receipt naming the competing forces, the deciding principle, the assumptions, and what would break the conclusion.
when_to_use
"Why did you pick that?", "how did you resolve this conflict", "which option should we choose", after a recommendation or analysis that the user will act on. For finished multi-file changes, use change-brief instead; it already contains the key insight.
# Reasoning receipt The aim is to teach a **reusable reasoning move**, not to justify yourself. Keep it to 3–6 lines, in plain language a smart newcomer would follow. ## For a recommendation or answer ``` Reasoning - Core question: <the real question, in one line> - Key insight: <the first-principles fact that decides it> - Assumes: <what must be true; the weakest assumption first> - Breaks if: <the condition that would flip the answer> - Rejected: <best alternative>, because <reason> ``` ## For a conflict ``` How the conflict was resolved - Force A: <wants x> because <underlying need> - Force B: <wants y> because <underlying need> - Deciding principle: <invariant or priority, e.g. "money correctness > latency", "the user's data is never lost"> - So: <resolution> · Cost accepted: <what we gave up> · Revisit if: <trigger> ``` Guidelines: - Name the *underlying need* behind each side, not just its position. Most conflicts dissolve once the needs are visible, and the resolution often satisfies both. - Prefer principles that transfer: invariants, constraints, costs of being wrong, reversibility, who bears the risk. - If the evidence was thin, say so: "Low confidence: based on one log line." - For merge conflicts: state what each branch intended (from commits or PRs), and which intent the merged result preserves.
Ver en GitHub