cognitive-fallacies-guard_skill

Overview

This skill detects and prevents visual misleads, cognitive biases, and data integrity failures in charts, dashboards, reports, and presentations. It focuses on actionable audits that identify where visual design or data choices could distort interpretation and recommends concrete fixes. Use it to ensure visual honesty and reduce user misinterpretation before publication.

How this skill works

The skill runs a three-step audit: scan for visual misleads (chartjunk, 3D effects, truncated axes), check for cognitive bias exploitation (framing, anchoring, cherry-picking), and verify data integrity (complete context, honest axes, fair comparisons). For each issue found it flags severity (critical/high/medium/low) and suggests a specific remediation. The workflow is fast (15–30 minutes) and designed for reviewers, analysts, and presenters.

When to use it

• Before publishing dashboards or reports to check for misleading visuals
• When users report frequent misinterpretation of charts or KPIs
• During design reviews to catch perceptual distortions and chartjunk
• To audit suspected cherry-picked or truncated data displays
• When verifying visualizations for compliance, transparency, or ethics

Best practices

• Start audits with a quick visual scan for chartjunk and 3D effects
• Confirm bar/line axes start and scale conventions; prefer zero-baseline for bars
• Use consistent scales for comparisons and avoid dual-axis tricks
• Show full context and denominators; cite sources and limitations
• Prefer 2D proportional encodings and clear labels over decorative styling

Example use cases

• Reviewing quarterly sales charts to ensure axes aren’t truncated
• Auditing a marketing dashboard for cherry-picked time ranges
• Checking investor slides for visual exaggeration using 3D or volume
• Evaluating a public report to verify source transparency and denominators
• Diagnosing why users misread a metric panel due to anchoring or framing