Source: https://github.com/aipoch/medical-research-skills

When to Use

• You converted a paper PDF to Markdown and need to verify that Results text matches figure legends (without inspecting the images).
• You want to detect missing figure references in the Results section (e.g., a legend describes an analysis not mentioned in text).
• You need to find numerical/label mismatches (e.g., group names, time points, units, n-values) between Results paragraphs and legends.
• You are preparing a revision and want an actionable discrepancy list down to the panel/sub-figure level.
• You need standardized outputs (Markdown report + CSV) for editorial or QA workflows.

Key Features

• Compares Results descriptions vs figure legend text using the PDF-to-Markdown source (including ## Page XX markers).
• Produces:
  • A Markdown consistency report (UTF-8).
  • A CSV issue list (UTF-8) with structured fields for tracking and revision.
• Enforces a template-based report format using assets/consistency_template.md.
• Uses a rule/checklist reference from references/guide.md.
• Text-only validation: does not read figure images and does not infer visual content.

Dependencies

• None (no external runtime dependencies specified).
• Input prerequisite (if starting from PDF): a PDF-to-Markdown conversion step (e.g., pdf-extract) must be completed before running this check.

Example Usage

Input

Place the converted full text Markdown in your working location (example: inputs/paper_fulltext.md). The file should include page headers like:

## Page 12
... Results text ...

Figure 3. ...
(A) ...
(B) ...

Run (conceptual workflow)

1. Read the full Markdown input (PDF conversion output).
2. Identify:
   • Result paragraphs describing findings.
   • Figure legend blocks (including panel labels such as A/B/C).
3. Compare legend statements against Results statements and record discrepancies.
4. Write outputs to outputs/:
   • outputs/consistency_report.md (UTF-8)
   • outputs/consistency_issues.csv (UTF-8)

Output files

outputs/consistency_issues.csv (UTF-8) columns:

Figure Number,Location/Reference,Issue Description,Suggested Revision,Priority

Notes:

• Location/Reference must contain only Page XX.
• Issues should be granular to the panel/sub-figure level when applicable.

outputs/consistency_report.md (UTF-8) must follow:

• Template: assets/consistency_template.md
• If no issues are found, write "None found" in the relevant sections.

Implementation Details

• Scope of comparison

  • Only compare main text Results descriptions and figure legend text present in the Markdown input.
  • Do not inspect images or infer information not explicitly stated in text.

• Rules and checklist

  • Follow the specific checking rules and required output points defined in:
    • references/guide.md

• Report formatting

  • The Markdown report must be generated by filling:
    • assets/consistency_template.md
  • Ensure all outputs are saved under:
    • outputs/ (within the skill directory)

• Granularity and actionability

  • Record discrepancies at the most specific level possible (e.g., Figure 2B vs Figure 2 overall).
  • Provide a concrete Suggested Revision whenever feasible (e.g., align terminology, correct numbers/units, add missing reference).

• Language

  • Default output language is Chinese.
  • If the user explicitly specifies another language, output in that language.

When Not to Use

• Do not proceed when required input files, identifiers, parameters, or context are missing — ask the user to provide them first.
• Do not assume capabilities beyond this skill's declared scope when the user requests external operations or inferences.
• Do not proceed without user confirmation when overwriting existing results, executing high-cost batch operations, or expanding task scope.

Required Inputs

Output Contract

• Primary output: Structured result or target file aligned with this skill's objective.
• Optional output: Intermediate check notes, issue list, supplementary suggestions, or generated file paths.
• Format requirement: Unless the user specifies otherwise, prefer stable, reviewable Markdown or JSON; if the skill's bundled script requires a fixed format, use that format.
• If partially complete: Must explicitly mark as PARTIAL and state which steps are completed and which remain.

Failure Handling

• Missing critical input: Explicitly state which fields, files, or identifiers are missing and pause.
• Script, template, or resource execution failure: Report the failing step, likely cause, and recovery suggestions — do not silently degrade.
• Partial completion only: Return the verified portion first, then list remaining blockers and suggested next steps.

User Checkpoints

• Before executing batch processing, overwriting files, long-running searches, or multi-stage generation, confirm scope and output format with the user.
• Before proceeding when a key judgment is ambiguous, evidence is insufficient, or the workflow is entering the next stage, confirm with the user.

Input Validation

This skill accepts requests that match the documented purpose of result-figure-consistencycheck and include enough context to complete the workflow safely.

Do not continue the workflow when the request is out of scope, missing a critical input, or would require unsupported assumptions. Instead respond:

result-figure-consistencycheck only handles its documented workflow. Please provide the missing required inputs or switch to a more suitable skill.

Quick Validation

• Check that key scripts, templates, or reference file paths this skill depends on exist.
• Check that the final output contains the core fields, sections, or files specified for this task.
• Check that results clearly mark assumptions, limitations, and incomplete items.