# Document Comparison

> Compare extracted documents against rubrics and reference documents with traceable evidence.

- Skill: `bionic-gpt/document-comparison` (Agent Skill)
- Install (CLI): `npx skillmds@latest add bionic-gpt/document-comparison`
- Raw SKILL.md: https://api.skillmd.com/api/skills/bionic-gpt/document-comparison/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: bionic-gpt (https://skillmd.com/u/bionic-gpt)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/bionic-gpt/document-comparison

---

# Document Comparison

Use this skill when a user asks to compare documents, validate a document against a rubric or policy, reconcile evidence across files, or identify gaps and risks.

## Workflow

1. Identify the governing rubric, checklist, policy, contract, or reference document and distinguish it from the evidence documents.
2. Read `/home/user/skills/structured-extraction/SKILL.md` and use that workflow to extract every relevant source before drawing conclusions.
3. Build a requirement-by-requirement matrix with the requirement, status, supporting document, precise location, and action needed.
4. Use exactly these statuses: **Pass**, **Fail**, **Partial**, and **Unknown**.
5. Reconcile conflicts explicitly. Keep source facts, conclusions, and recommendations separate.

## Quality checks

- Every governing requirement appears exactly once in the matrix.
- Every Pass, Fail, or Partial result cites explicit evidence and a location.
- An omitted clause is Unknown unless another document supplies explicit evidence.
- Conflicting documents are called out rather than silently resolved.