What kind of language does this skill prohibit?

It bans unsubstantiated superlatives (e.g., 'industry-leading', 'optimal') and fabricated metrics in favor of factual, measurable observations.

Can this skill help with software project estimations?

Yes, it mandates that time and effort estimates be derived from file counts and measured code complexity rather than arbitrary guesses.

How does it handle information that cannot be verified?

The skill instructs Claude to use explicit uncertainty markers such as 'requires investigation' or 'unable to confirm' rather than generating a plausible guess.

How does the Anti-Fabrication skill prevent AI hallucinations?

It requires Claude to execute specific system tools like Read, Grep, or Bash to verify the state of a project before making any factual claims or assertions.

Anti-Fabrication & Fact Verification

Name: Anti-Fabrication & Fact Verification
Author: vinnie357

byvinnie357

学习与文档

Enforces factual accuracy in AI outputs by requiring tool-based verification and prohibiting unsubstantiated claims or metrics.

关于

The Anti-Fabrication skill provides a rigorous framework for maintaining factual integrity by mandating evidence-based responses. It forces Claude to execute system tools—such as file reading, grep searches, and bash commands—before making assertions about codebases, system performance, or project timelines. By replacing vague superlatives and fabricated metrics with verifiable data, this skill ensures that documentation, research, and technical reports remain precise, transparent about uncertainties, and free from common AI hallucinations or 'plausible-sounding' fabrications.

主要功能

Mandatory tool-based verification (Read, Glob, Bash, Grep) for all assertions
Prohibition of unverified superlatives like 'cutting-edge' or 'robust'
Strict validation workflows for technical documentation and research
0 GitHub stars
Data-backed time and effort estimation based on measurable complexity
Explicit uncertainty marking for unverified or missing information

使用场景

Generating high-precision technical documentation and API specifications
Providing evidence-based project estimates and effort assessments
Reporting validated test results and system performance benchmarks

关于

主要功能

Mandatory tool-based verification (Read, Glob, Bash, Grep) for all assertions
Prohibition of unverified superlatives like 'cutting-edge' or 'robust'
Strict validation workflows for technical documentation and research
0 GitHub stars
Data-backed time and effort estimation based on measurable complexity
Explicit uncertainty marking for unverified or missing information

使用场景

Generating high-precision technical documentation and API specifications
Providing evidence-based project estimates and effort assessments
Reporting validated test results and system performance benchmarks