Back to skill

Security audit

公众号内容复盘

Security checks for vulnerabilities and agentic risk

Overview

The skill is rated suspicious because its commands use an automatically installed, unpinned latest CLI package, creating supply-chain risk; the scanner coverage is partial.

Consider waiting until the CLI is pinned to a reviewed version before installing, especially if the environment contains credentials or sensitive files. The scan covered one Markdown component but was marked partial because two local references could not be resolved, so the evidence is incomplete.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (10)

Credential Access

High
Category
Privilege Escalation
Confidence
70% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · SKILL.md (reported line 27)May include surrounding context.

md
` 与来源 `environment` / `file` / `null`,不输出 Key。安装或升级时,已有凭证可读、没有已知鉴权失败且用户未明确要求更换 Key,就复用原来源继续原任务,不重新登录、不迁移或改写已有凭证;命令、权限或文件格式错误按实际错误处理,不一律判为没配置。已认证的 hosted MCP 沿用当前连接鉴权,不要求另配本地 CLI。
- 仅首次配置、凭证缺失或失效,或用户明确要求更换 Key 时才录入:用户已在安装提示词或当前对话中提供 Key 并授权配置时,用 `npx -y socialdatax-skills@latest auth login --stdin` 接收这枚 Key 的标准输入,由 CLI 长期保存;不要求用户重复手动录入,不把 Key 作为 CLI 参数或写入临时脚本文件。没有提供 Key 时,引导用户从上述官网获取,在本地终端运行 `npx -y socialdatax-skills@latest auth login`,按隐藏输入提示填写,不主动要求其发到对话。Key 保存到当前执行用户主目录的 `.config/socialdatax/credentials.json`;Unix 权限 600,Windows 当前用户私有 ACL。不要在回复、输出日志、截图、命令示例、Skill 或版本库再次展示 Key;Agent 不读取或展示凭证文件内容。
- CLI 每次调用都会读取该私有用户文件,不依赖客户端加载 Shell 配置,不需要改客户端或系统设置。非空 `SOCIALDATAX_API_KEY` 优先,既有旧兼容变量其次,最后读取文件;`source=environment` 时保存文件不会覆盖生效的环境凭证;保存成功不等于本次所选 Key 已生效。CLI 提示不同 Key 冲突时,先说明实际来源,不擅自删除或改写既有环境配置。鉴权失败先核对实际来源和账号,再由用户本地修正,保留其它配置。
- 完成保存后,在实际业务命令的同一执行环境重新运行 `auth status`。false 时停止业务调用,核对操作系统用户、主目录与文件访问权限;true 只说明凭证可读,鉴权与结果仍以原业务请求返回为准,不为配置检查额外调用收费接口。
- 远程或容器中的 CLI 读取远程或容器用户的配置,本机保存不会自动传入。具备持久主目

MCP server referenced without pinned version: 'npx -y socialdatax-skills'.

Medium
Category
MCP Rug Pull
Confidence
99% confidence
Finding

The authentication-status command executes socialdatax-skills@latest through npx with automatic installation approval. Although this is a CLI rather than an MCP server, the unpinned execution creates a supply-chain exposure: a compromised future release could execute with the agent's permissions and access available credentials, regardless of the command's stated read-only purpose.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx -y socialdatax-skills'.

Medium
Category
MCP Rug Pull
Confidence
99% confidence
Finding

The interactive login example also runs socialdatax-skills@latest without a fixed version. Hidden terminal input protects against casual disclosure, but a compromised CLI release could capture the entered key or execute other actions as the current user.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx -y socialdatax-skills'.

Medium
Category
MCP Rug Pull
Confidence
99% confidence
Finding

The stdin-based login command executes a mutable @latest package while directly supplying an API key to that executable. Standard input avoids exposing the key in command arguments, but it cannot protect the key from a compromised package receiving it.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx -y socialdatax-skills'.

Medium
Category
MCP Rug Pull
Confidence
99% confidence
Finding

The article-fetch example automatically executes the current @latest npm release rather than a reviewed fixed version. A compromised release could access the API key and local resources; the intended read-only API operation does not constrain the local executable.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx -y socialdatax-skills'.

Medium
Category
MCP Rug Pull
Confidence
99% confidence
Finding

The account-article listing command uses npx -y with the mutable @latest tag. Its pagination limits control legitimate API usage, but do not prevent arbitrary local execution or credential theft if the resolved package is compromised.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx -y socialdatax-skills'.

Medium
Category
MCP Rug Pull
Confidence
99% confidence
Finding

The statistics command runs an unpinned npm release with automatic installation approval. This exposes the execution environment and configured API key to any malicious code introduced into the selected release, even though statistics retrieval itself is read-only.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx -y socialdatax-skills'.

Medium
Category
MCP Rug Pull
Confidence
99% confidence
Finding

The account-information command executes socialdatax-skills@latest rather than a fixed reviewed release. The scanner incorrectly describes it as an MCP server reference, but the underlying supply-chain risk remains because npm-delivered code runs with local user permissions.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx -y socialdatax-skills'.

Medium
Category
MCP Rug Pull
Confidence
99% confidence
Finding

The comments command automatically resolves and executes the latest CLI package. Comment-count limits reduce normal request volume, but do not mitigate malicious package code accessing credentials or modifying files in the execution environment.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx -y socialdatax-skills'.

Medium
Category
MCP Rug Pull
Confidence
99% confidence
Finding

The replies command repeats the same unpinned @latest execution pattern as the other CLI examples. A compromised resolved release could run arbitrary code with the agent's permissions, so the skill's read-only business scope does not eliminate the supply-chain exposure.

Content

No source excerpt is available for this finding.