Comparison Report

Guidelines for producing comparison reports between two or more options.

Output Structure

1. Criteria: List the dimensions used for comparison (e.g. performance, ecosystem, learning curve).
2. Comparison Table: Side-by-side comparison. Rows = criteria, columns = options. Use inline [Source] for metrics when citing.
3. Pros and Cons: Per-option strengths and weaknesses.
4. Recommendation: Clear conclusion with rationale. Map to scenarios when appropriate.
5. References: Required. ## References section listing all sources (artifact names, URLs, or "Official documentation" for general knowledge). Quantitative claims in the table must have attribution.

Table Format

| Criterion    | Option A | Option B | Option C |
|--------------|----------|----------|----------|
| Performance  | High     | Medium   | High     |
| Ecosystem    | Mature   | Growing | Mature   |
| Learning     | Moderate | Easy    | Steep    |

• Use consistent scale or labels (e.g. High/Medium/Low, or specific metrics).
• Add a brief "Notes" row if needed for context.
• When metrics are numeric, include units (e.g. "10ms", "50 req/s").

Pros/Cons Format

Option A

• Pros: ...
• Cons: ...

Option B

• Pros: ...

• Cons: ...

• Aggregate pros/cons from input artifacts; avoid duplication.

• Prioritize by importance when many items.

Recommendation

• State the recommended option clearly.
• Include "when to use" or "best for" scenarios.
• Support with evidence from the comparison.
• If no clear winner: present trade-offs and suggest "depends on X".

Input Artifacts

• Use ReadArtifact to get research outputs for each option (e.g. task_1, task_2, task_3).
• Combine into unified comparison. Ensure criteria align across sources.
• If artifacts use different criteria, create a unified criteria set and fill from each source.

Output Format

• Markdown for human-readable reports.
• JSON for structured output when spec requires:

  {
    "options": [{"name": "A", "pros": [...], "cons": [...]}],
    "comparison": {"criterion": {"A": "value", "B": "value"}},
    "recommendation": {"option": "A", "rationale": "..."}
  }

Quality Checklist

• All options from input artifacts included
• Criteria cover key dimensions
• Table is complete (no empty cells without reason)
• Recommendation supported by evidence
• ## References section present; quantitative metrics in table have source attribution