name: contradiction-analysis
description: >
Detect contradictions and ambiguities in biomedical evidence using the
BioTeam-AI W6 Ambiguity Engine. Analyzes claims across papers, scores
evidence with RCMXT, and generates hypothesis-driven research directions
from identified contradictions.
metadata:
author: bioteam-ai
version: "1.0"
domain: biomedical-research
allowed-tools: Bash(python:*) Read
Contradiction Analysis Skill
When to Activate
- User asks to "find contradictions" in a set of papers or findings
- User wants to assess ambiguity or conflicting evidence
- User asks about research gaps that stem from contradictory results
Usage
cd backend && uv run python -m app.skills.w6_cli --topic "your research topic" --budget 2.0
Pipeline
- Gather evidence from Lab KB and prior W1 literature reviews
- Run ContradictionDetector across all claim pairs
- Score each contradiction with RCMXT framework
- Generate hypotheses explaining contradictions
- Produce structured report with action items
Output
- Contradiction matrix with confidence scores
- RCMXT evidence scores per conflicting claim
- Generated hypotheses for further investigation
- Recommended experimental approaches to resolve contradictions