Auto-Skiller Compare Options
This skill allows the agent to systematically measure, compare, and optimize choices to ensure the objectively best available path is selected with documented rationale.
Steps
- Enumerate all options. List every available option clearly. Do not pre-filter based on assumption.
- Define evaluation criteria. Identify the dimensions that matter (e.g., cost, speed, risk, compatibility, reversibility). Weight each criterion if applicable.
- Score each option per criterion. Use a consistent scale (e.g., 1–5) and apply it neutrally across all options.
- Build the comparison matrix. Organize scores into a table for clear visual comparison.
- Identify the winner. Select the option with the highest weighted score, noting any exceptions (e.g., a hard constraint that disqualifies a leading option).
- Document trade-offs. Explicitly state what is sacrificed with the winning choice.
- Output the recommendation with the matrix, winner, rationale, and trade-offs.
Output Format
benchmark_result:
criteria: ["criterion_1", "criterion_2"]
matrix:
option_a: {criterion_1: 4, criterion_2: 3}
option_b: {criterion_1: 2, criterion_2: 5}
winner: "option_b"
rationale: "<why this wins>"
trade_offs:
- "<what is sacrificed by choosing option_b>"
Notes
- Always make criteria explicit before scoring — implicit criteria lead to biased outcomes.
- If options have hard disqualifying constraints, flag them before scoring.