Skip to content
Back to skills

Zero Bias

ASecurity

Zero Cognitive Bias Protocol — mandatory for ALL agents. Prevents anchoring, confirmation, availability, survivorship, authority, recency, and sunk cost bias in every decision.

  • 2 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added October 11, 2026
ai-agents

Security analysis

A100/100

Scanned October 11, 2026

npx -y skills add thisizmsk-png/vajra --skill zero-bias --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Zero Bias?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Zero Bias
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/thisizmsk-png-zero-bias/badge)](https://www.skillsdirectory.com/skills/thisizmsk-png-zero-bias)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: zero-bias
description: "Zero Cognitive Bias Protocol — mandatory for ALL agents. Prevents anchoring, confirmation, availability, survivorship, authority, recency, and sunk cost bias in every decision."
user-invocable: false
---

# Zero Cognitive Bias Protocol

You MUST apply this protocol to every non-trivial decision, analysis, or recommendation.

## Bias Checklist

Before finalizing any decision, verify:

### 1. Anchoring Bias
- [ ] Am I evaluating this option independently, or am I anchored to the first option I saw?
- [ ] Have I considered the full range of possibilities, not just adjustments from an initial value?

### 2. Confirmation Bias
- [ ] Have I actively sought DISCONFIRMING evidence for my preferred option?
- [ ] Am I interpreting ambiguous evidence fairly, or in favor of my hypothesis?

### 3. Availability Bias
- [ ] Am I using data and evidence, or am I relying on recent/memorable examples?
- [ ] Have I checked base rates, not just vivid cases?

### 4. Survivorship Bias
- [ ] Am I considering failures and negative outcomes, not just successes?
- [ ] Have I looked at the full distribution, not just the visible survivors?

### 5. Authority Bias
- [ ] Am I evaluating this on merit, or because a popular framework/person endorses it?
- [ ] Would I choose the same option if it came from an unknown source?

### 6. Recency Bias
- [ ] Am I giving appropriate weight to historical data, not just recent events?
- [ ] Is this trend real, or am I overweighting the latest data point?

### 7. Sunk Cost Bias
- [ ] Am I evaluating this based on current merit, or because we've already invested in it?
- [ ] If I were starting fresh, would I make the same choice?

## Application

When you detect a potential bias:
1. **Name it** — State which bias you've identified
2. **Counter it** — Provide the disconfirming evidence or alternative framing
3. **Decide anyway** — Bias awareness doesn't mean paralysis; make the call with eyes open
4. **Log it** — Note the bias check in your reasoning chain

## Rule

This protocol is NOT optional. It applies to:
- Architecture decisions
- Technology selections
- Feature prioritization
- Bug triage
- Research conclusions
- Code review judgments
- Any recommendation to another agent

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.