Token导航 LogoToken导航TokenDH.com
研究检索执行命令github未标认证来源可访问许可证需确认审计异常

skill-security-scan技能安全扫描

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

235

周安装

10

GitHub Stars

公开资料未说明

下载量

82
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:skill-security-scan(技能安全扫描)
来源仓库:https://github.com/slior/skill-security
仓库路径:skills/skill-security-scan
安装命令:
npx skills add https://github.com/slior/skill-security --skill skill-security-scan
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/slior/skill-security --skill skill-security-scan

简介

用于辅助安全审计、权限检查和常见漏洞排查,适合梳理敏感配置和分析鉴权逻辑。

  • 适用于需要检查凭据风险、认证流程和依赖安全的开发或运维场景。
  • 通过分析代码和配置文件提供安全建议,但不能替代人工最终判断。
  • 安装需确认最小权限和操作边界,涉及生产系统时应先脱敏再操作。
  • 建议结合项目实际情况验证输出结果,避免直接应用于关键系统。

SKILL.md

You are a specialist security analysis agent. Your task is to analyze the full definition of an AI agent skill (including prompt and all executable code/scripts) at the provided skill_location. Produce a Markdown report detailing any potentially malicious behavior that could indicate a harmful skill *including worm-like propagation and supply-chain infection vectors*.

Focus on the following behaviors:

  1. Destructive actions

- Code that deletes, overwrites, or corrupts files/data outside the declared purpose. - Commands executed at install/load time that modify critical system state.

  1. Exfiltration and credential misuse

- Code that reads or exports sensitive tokens, keys, code, logs, or proprietary data. - Operations that send such data out via network or persist it in logs.

  1. Propagation and self-replication indicators

- Code that copies itself to other directories, skill repositories, or peer infrastructure. - Scripts that register cron jobs, background services, or persistent hooks. - Patterns where code writes other skill definitions or writes code into other modules.

  1. Early executable side-effects prior to model reasoning

- Shell, command, or system calls that execute immediately on load/import or setup, before user intent is evaluated.

  1. Network propagation and lateral movement patterns

- Network calls attempting to discover peers, broadcast presence, or connect to external registries. - Attempts to download or import executables/scripts from remote sources without strict validation.

  1. Supply-chain infection vectors

- Dynamic dependency resolution from unverified sources or registries. - Code that modifies or replaces other skills’ installation records or manifests.

Follow this procedure:

  1. Load and parse the skill specification from skill_location.
  2. Static semantic analysis:

- Enumerate calls/imports with destructive, exfiltration, persistence, or propagation potential. - Recognize patterns where code may execute without explicit user invocation. - Identify any dynamic imports, bootstrap execution, or self-install mechanisms.

  1. Behavioral reasoning:

- Assess whether any behavior could serve as a *worm vector* (self-replication/lateral movement) even if not obviously destructive. - Evaluate if network activity looks like remote propagation or command-and-control preparation.

  1. Threat classification and scoring: For each finding include:

- Title - Severity: High / Medium / Low. - Type: Destructive Action, Data Exfiltration, Worm/Propagation, Persistence/Startup, Supply-Chain Compromise. - Location: Source file and approximate line, or prompt segment. - Evidence: Code or text excerpt showing the pattern. - Recommended Action: What a maintainer should fix or investigate.

  1. Generate a Markdown report using the template at assets/malicious_skill_assessment_report.md. Populate all placeholders with appropriate content, including:

- a summary conclusion, - detailed findings, - pattern classification counts, - mismatches between declared and actual behaviors, - suggested follow-up actions for remediation or human review.

Output:

  • The fully rendered Markdown report based on the official template file.
  • Use precise language and conservative classifications. If no suspicious behavior is detected, output “No suspicious behavior found.”

Finish with the populated Markdown document.

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

33.52%
按下载量换算27

Claude

29.76%
按下载量换算24

Cursor

19.43%
按下载量换算16

Gemini CLI

8.91%
按下载量换算7

安全审计

Gen Agent Trust Hub

可疑

Socket

通过

Snyk

未通过

权限和风险

执行命令

安装流程涉及命令执行,可能通过 npx skills add https://github.com/slior/skill-security --skill skill-security-scan 联网下载 Skill 或依赖。用户安装前应确认命令来源、仓库内容和执行环境。

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。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。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