comparison-matrix-builder

name	comparison_matrix_builder
description	Build a structured comparison matrix from paper cards so the drafting stage can reason about methods, datasets, benchmarks, and limitations without rereading every abstract.
backend	local_function
backend_ref	fn:comparison_matrix_builder
default_agent	analyst
output_artifact_type	comparison_matrix
visibility	both
tags	["analysis","comparison","papers","survey","matrix"]
input_schema	{"paper_cards":{"type":"array"},"compare_dimensions":{"type":"array"},"format":{"type":"string"}}

Comparison Matrix Builder

Use this skill before drafting whenever the workflow needs a stable, auditable summary of the paper set.

Workflow

1. Normalize each paper card into a row keyed by paper title.
2. Collect methods, datasets, benchmark signals, and limitations from extracted fields.
3. Mark missing evidence instead of hallucinating it.
4. Return both machine-friendly rows and a human-readable table so downstream prompts can consume either form.

Output Contract

• matrix.rows: structured rows with one paper per row.
• matrix.table_text: markdown-ready table string for debugging or prompt injection.
• missing_fields: explicit reminders about weak evidence coverage.