confidence-honesty

Require numeric confidence percentages with justification and gap analysis for conclusions.

334|43|Updated Nov 14, 2025

Search Tags:#validation #analysis #evidence #uncertainty #confidence #decision-support

Ready to run

npx skills add https://github.com/NTCoding/claude-skillz --skill confidence-honesty-ntcoding

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: confidence-honesty
Source: https://github.com/NTCoding/claude-skillz/tree/main/confidence-honesty
Command: npx skills add https://github.com/NTCoding/claude-skillz --skill confidence-honesty-ntcoding

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill enforces explicit confidence reporting before making conclusions by requiring a quantified confidence percentage, a justification, and a gap analysis as new evidence emerges, reducing overconfident or unvalidated claims.

Core Features & Use Cases

Express confidence as a percentage: Every conclusion includes a numeric confidence level with calibrated interpretation.
Evidence and gaps: Automatically attach supporting evidence and clearly state what remains uncertain.
Self-validation prompts: Mandate "Why not 100%?" reasoning and propose next steps to gather missing data.
Use cases: Ideal for research reports, investigations, and decision-support where stakeholder trust matters.

Quick Start

Run the Confidence Honesty protocol before presenting any conclusions. Example: "Apply Confidence Honesty: state 72% confidence, list supporting evidence, explain gaps, and justify why not 100%."

Frequently Asked Questions about confidence-honesty

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I add explicit confidence percentages to decision-support conclusions?▼

To add explicit confidence percentages to decision-support conclusions, you enforce a numeric confidence level, provide supporting evidence, and explain existing gaps to reduce overconfident claims.

What is confidence reporting and when do I need it for analysis?▼

Confidence reporting is the process of quantifying uncertainty with a percentage for every conclusion. You need it for analyses, investigations, and decision-support tasks where stakeholder trust matters.

How do I validate analysis uncertainty and justify why a conclusion is not 100% certain?▼

You validate analysis uncertainty by applying self-validation prompts that mandate 'Why not 100%' reasoning. This requires identifying evidence gaps and proposing next steps to gather missing data.

Does decision-support validation work without structured evidence gap identification?▼

Decision-support validation requires structured evidence gap identification to enforce explicit confidence reporting. Without identifying gaps and justifying confidence levels, conclusions risk being overconfident or unvalidated.

What is the best way to structure validation rules for research reports?▼

The best way to structure validation rules for research reports is to require a quantified confidence percentage, list supporting evidence, explain remaining uncertainties, and justify why the confidence level is not absolute.

When should I not use confidence percentages for conclusions?▼

You should not use confidence percentages for conclusions when uncertainty does not matter or when stakeholder trust is not a factor, as the structured validation process adds overhead without proportional value.