Skip to main content

official-verifier

判断网站或项目仓库是否为官方项目,通过多种验证手段并以概率形式返回可信度评估。Judges whether a website or project repository is official, evaluating authenticity through multiple verification methods and returning a probability-based assessment.

Datos de origen

Repositorio
Kernel8901/ai-agent-skills-classification
Última actividad en el origen
4 de abril de 2026 a las 15:26
Idioma detectado de SKILL.md
Varios idiomas
Estrellas
5
Forks
1

Opciones de instalación

De forma predeterminada está seleccionado el prompt que primero revisa el origen. Puedes cambiar a un comando directo o descargar una copia local.

Revisa los archivos de origen

Lee SKILL.md y los archivos complementarios que muestra SkillsMP antes de decidir si quieres instalarlo.

Explorador de archivos
10 archivos

Mostrando SKILL.md

SKILL.md
Instrucciones de origen · Vista previa de solo lectura
name
official_verifier
description
判断网站或项目仓库是否为官方项目,通过多种验证手段并以概率形式返回可信度评估。Judges whether a website or project repository is official, evaluating authenticity through multiple verification methods and returning a probability-based assessment.
# 官方项目验证器技能 (Official Project Verifier Skill) ## 中文描述 此技能用于判断网站或项目仓库是否为官方项目,通过多种验证手段并以概率形式返回可信度评估。 ## English Description This skill is used to determine whether a website or project repository is an official project, assessing authenticity through multiple verification methods and returning a probability-based credibility evaluation. ## 使用方法 (Usage) 当用户询问某个项目或网站是否为官方项目时,使用此技能进行分析。技能将通过以下维度评估项目真实性: - GitHub Star数量 (GitHub Star Count) - 开发者认证 (Developer Verification) - 信息一致性 (Information Consistency) - 第三方引用 (Third-party References) - 项目活跃度 (Project Activity) - 安全指标 (Security Indicators) ## 输出格式 (Output Format) 返回JSON格式的结果,包含以下字段: - probability: 数值型,0-100,表示官方项目的可能性百分比 (Numeric value, 0-100, indicating percentage probability of being official) - confidence: 字符串,"very_high", "high", "medium", "low", "very_low" (String, "very_high", "high", "medium", "low", "very_low") - reasons: 数组,列出支持判断的具体理由 (Array, listing specific reasons supporting the judgment) - warnings: 数组,列出潜在风险点 (Array, listing potential risk points) - score_breakdown: 对象,包含各维度得分详情 (Object, containing detailed scores for each dimension) ## 工具使用 (Tool Usage) 使用内置的verify.js脚本执行验证逻辑,该脚本实现了多维度评分系统。 Uses the built-in verify.js script to execute verification logic, implementing a multi-dimensional scoring system.
Ver en GitHub