返回 Skill 列表
extension
分类: AI Agent 能力无需 API Key

compare-options

多位专家对选项进行比较和排序,并进行结构化审议。主动激活用于:(1)在备选方案之间做出决策,(2)权衡分析,(3)选项评估,(4)供应商/工具比较,(5)策略选择。触发词:“比较选项”,“评估备选方案”,“对选项排名”,“我应该选择哪一个”,“比较这些”,“权衡分析”,“在...之间做决定”

person作者: jakexiaohubgithub

Compare Options

Multi-expert comparison and ranking of options using structured deliberation.

When to Use

Use this skill when you need to:

  • Compare multiple alternatives systematically
  • Evaluate trade-offs between options
  • Get diverse expert perspectives on a decision
  • Rank options with transparent criteria
  • Make decisions with documented rationale

Workflow

Invoke the expert-panel-deliberation skill for:

Options to compare: $ARGUMENTS

Default Parameters

| Parameter | Value | Rationale | |-----------|-------|-----------| | panel_size | 5 | Diverse perspectives without analysis paralysis | | output_format | ranking | Produces prioritized list with scores | | deliberation_depth | standard | Balanced thoroughness | | include_challenger | true | Include devil's advocate perspective | | consensus_required | false | Allow minority opinions |

Expert Panel Assembly

The panel is dynamically composed based on the decision domain:

For Technical Decisions:

  • Solution Architect
  • Performance Specialist
  • Security Expert
  • DevOps/Operations
  • End User Advocate

For Business Decisions:

  • Strategic Planner
  • Financial Analyst
  • Market Expert
  • Operations Lead
  • Customer Representative

For Product Decisions:

  • Product Strategist
  • UX Designer
  • Engineering Lead
  • Business Analyst
  • Customer Advocate

Deliberation Process

Round 1: Independent Assessment

Each expert evaluates all options against criteria:

  • Scores each option (1-10) on their specialty dimensions
  • Documents key strengths and weaknesses
  • Identifies critical risks or blockers

Round 2: Cross-Examination

Experts challenge each other's assessments:

  • Surface hidden assumptions
  • Probe trade-off implications
  • Identify overlooked factors

Round 3: Synthesis

Panel converges on:

  • Weighted ranking of options
  • Key trade-offs documented
  • Recommended choice with rationale
  • Conditions under which ranking changes

Evaluation Criteria Framework

Standard criteria (customized to domain):

| Criterion | Weight | Description | |-----------|--------|-------------| | Effectiveness | 25% | How well does it solve the core problem? | | Cost/Effort | 20% | Total cost of ownership, implementation effort | | Risk | 20% | What can go wrong? Likelihood and impact | | Scalability | 15% | Does it grow with needs? | | Reversibility | 10% | Can we change course if needed? | | Time to Value | 10% | How quickly do we see benefits? |

Output Format

The comparison produces:

<option-comparison>
  <header>
    <id>[unique identifier]</id>
    <decision_context>[What we're deciding]</decision_context>
    <options_count>[number of options]</options_count>
  </header>

  <options>
    <option id="[A|B|C|...]" name="[option name]">
      <description>[Brief description]</description>
    </option>
  </options>

  <expert-panel>
    <expert role="[role]" specialty="[area]">
      <scores>
        <score option="A" value="[1-10]" rationale="[brief reason]"/>
        <score option="B" value="[1-10]" rationale="[brief reason]"/>
      </scores>
      <recommendation>[Expert's top pick and why]</recommendation>
    </expert>
    <!-- ... more experts ... -->
  </expert-panel>

  <deliberation-summary>
    <consensus_points>
      <point>[What experts agreed on]</point>
    </consensus_points>
    <debate_points>
      <point>[Where experts disagreed and resolution]</point>
    </debate_points>
  </deliberation-summary>

  <final-ranking>
    <rank position="1" option="[option]" score="[weighted score]">
      <rationale>[Why this is ranked first]</rationale>
      <caveats>[When this might not be the best choice]</caveats>
    </rank>
    <rank position="2" option="[option]" score="[weighted score]">
      <rationale>[Why second]</rationale>
      <when_preferred>[Conditions where this beats #1]</when_preferred>
    </rank>
    <!-- ... more ranks ... -->
  </final-ranking>

  <trade-off-matrix>
    <trade-off>
      <options>[Option A vs Option B]</options>
      <dimension>[What you trade off]</dimension>
      <analysis>[Detailed trade-off analysis]</analysis>
    </trade-off>
  </trade-off-matrix>

  <recommendation>
    <primary>[Top recommended option]</primary>
    <confidence>[high|medium|low]</confidence>
    <conditions>[When this recommendation applies]</conditions>
    <alternative>[When to choose differently]</alternative>
  </recommendation>

  <next-steps>
    <step>[Concrete action to proceed]</step>
  </next-steps>
</option-comparison>

Quality Gates

  • [ ] All options fairly represented
  • [ ] Criteria weights appropriate for context
  • [ ] Each expert provided independent assessment
  • [ ] Trade-offs explicitly documented
  • [ ] Minority opinions captured
  • [ ] Recommendation includes confidence and conditions
  • [ ] Next steps are actionable

Related Skills

After comparing options, consider:

  • Run /research-brief to deep-dive on top choice
  • Run /write-howto for implementation guide
  • Run /evaluate-schema if choice involves data design
  • Run /optimize-prompt if choice involves AI/prompting