Agent Skills and Tool Orchestration | Comparative Decision Prompt

What this prompt helps you do

Compare candidate options against one consistent decision framework. The output records evidence, weights, risks, and sensitivity checks so a human can review the recommendation.

Prompt text

You are a multi-agent and tool-orchestration architect. Compare [Option A], [Option B], and [Option C] for the objective: assign research, analysis, content, execution, and validation to specialized skills and agents. Inputs: [Overall objective], [Subtasks], [Tool capabilities], [Shared context], [Dependencies and failure modes]. First define five to eight non-overlapping evaluation dimensions and weights, and explain the rationale for the weights. Then build a scoring matrix in which every score is tied to a fact, data point, or explicit assumption. Provide the best choice, conditions under which it is best, irreversible risks, the lowest-cost validation method, and Agent and skill responsibilities, Invocation sequence, Input-output contracts, Shared state, Fallbacks and evaluation. Do not invent missing information to force a conclusion. Special requirement: Each agent should own a narrowly testable responsibility. Avoid circular calls, context contamination, and overlapping ownership.