ClawHub
Back to skill

Security audit

小红书博主内容研究

Security checks for vulnerabilities and agentic risk

Overview

The skill provides a coherently scoped creator-research workflow with no artifact evidence of hidden or unrelated actions.

Use a dedicated SocialDataX key stored through your client's secret settings, and submit only the account identifier or profile link needed. Queries use a third-party service and may consume paid credits. Prefer a reviewed, pinned CLI version and explicit page or item limits. This review covers the bundled instructions; the external CLI implementation and publisher identity were not independently verified.

A.I.G

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

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

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

Medium
Category
MCP Rug Pull
Confidence
98% confidence
Finding

Although this command invokes a direct CLI rather than an MCP server, it executes the mutable socialdatax-skills@latest package with automatic confirmation. A compromised publisher account or malicious release could execute code with the agent process's permissions and access its API key and other accessible secrets; the skill's read-only description does not constrain that code.

Content

No source excerpt is available for this finding.

The lookup workflow lacks a privacy notice that user-supplied account identifiers, profile URLs, or share text are sent to SocialDataX.

Medium
Category
Output Handling
Confidence
91% confidence
Finding

The skill instructs the agent to submit user-provided account identifiers or profile/share text through the SocialDataX CLI, but does not explicitly disclose this third-party data transfer. The read-only safety statement addresses account modifications and local browser access, not the privacy implications of sending lookup inputs to an external service.

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
98% confidence
Finding

The profile-URL variant repeats the same unpinned execution of socialdatax-skills@latest through npx -y. This creates the same supply-chain exposure as the preceding command, rather than a separate attack mechanism: a malicious release could access credentials or modify files despite the intended read-only API workflow.

Content

No source excerpt is available for this finding.