Token导航 LogoToken导航TokenDH.com
待分类只读github未标认证来源可访问许可证需确认审计通过

security-authentication安全认证

Agent Skill

用于辅助安全审计、权限检查、凭据风险、认证流程和常见漏洞排查。它适合让 Agent 梳理敏感配置、检查依赖风险、分析鉴权逻辑或生成安全复核清单。使用时不能把工具输出直接当最终结论,涉及密钥、令牌、用户数据或生产系统时,应先确认最小权限、脱敏方式和操作边界。

总安装

216

周安装

9

GitHub Stars

5

下载量

72
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:security-authentication(安全认证)
来源仓库:https://github.com/kentoshimizu/sw-agent-skills
仓库路径:skills/security-authentication
安装命令:
npx skills add https://github.com/kentoshimizu/sw-agent-skills --skill security-authentication
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/kentoshimizu/sw-agent-skills --skill security-authentication

简介

用于安全认证相关任务支持。security-authentication 属于待分类类 Skill,可作为该场景下的辅助能力补充。

  • 适合检查凭据风险和认证流程。适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。
  • 可生成安全复核清单或依赖风险报告。
  • 安装方式:通过 npx 从 GitHub 仓库添加技能。
  • 涉及生产系统时应确认最小权限。

SKILL.md

Security Authentication

Overview

Use this skill to design and review authentication flows that resist account takeover while preserving acceptable user friction.

Scope Boundaries

  • Authentication factors, login flows, or account-recovery behavior are being introduced or changed.
  • Session management (cookie/token TTL, refresh policy, revocation) needs to be defined.
  • Risk-based controls (MFA, step-up auth, suspicious login handling) are required.

Templates And Assets

  • Authentication assurance matrix:

- assets/auth-assurance-matrix-template.md

Inputs To Gather

  • Identity sources and trust level requirements (internal users, external users, federated identities).
  • Threat assumptions (credential stuffing, phishing, token theft, session hijacking).
  • Regulatory and product constraints (MFA mandates, session timeout policy, UX limits).
  • Operational constraints (IdP availability, incident response expectations, observability baseline).

Deliverables

  • Authentication flow map for primary login, re-auth, and recovery paths.
  • Credential and token/session policy (issuance, storage, rotation, revocation, expiry).
  • Control matrix for anti-abuse protections and detection signals.
  • Residual risk list with owners and verification checkpoints.

Workflow

  1. Define assurance targets by action sensitivity using assets/auth-assurance-matrix-template.md.
  2. Select factor strategy (password, passkey, OTP, federated SSO) using attacker capability and usability constraints.
  3. Design session/token lifecycle with explicit expiry, refresh, revocation, and device binding rules.
  4. Add anti-automation and abuse controls for login and recovery endpoints.
  5. Specify fallback and lockout policy that avoids permanent user denial while blocking attacker persistence.
  6. Define telemetry for login success/failure, suspicious patterns, and step-up triggers.
  7. Validate flows with negative scenarios: replay, stolen token use, brute-force, and recovery abuse.

Quality Standard

  • Every sensitive action has a declared required assurance level.
  • Session/token invalidation behavior is explicit and testable.
  • Recovery flow is at least as strong as primary authentication assurance.
  • Audit signals are actionable for incident triage.

Failure Conditions

  • Stop when account recovery can bypass primary assurance guarantees.
  • Stop when token/session revocation behavior is undefined.
  • Escalate when control strength cannot meet required risk level.

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

35.52%
按下载量换算26

Claude

29.87%
按下载量换算22

Cursor

19.68%
按下载量换算14

Gemini CLI

9.02%
按下载量换算6

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills