ClawHub
Back to skill

Security audit

Evernote Yinxiang

Security checks for vulnerabilities and agentic risk

Overview

This Evernote/Yinxiang skill is mostly purpose-aligned, but it asks for full note-account access while its dependency and API documentation does not match the included code.

Install only if you are comfortable granting a full-access Yinxiang/Evernote developer token. Protect the .env file, review any search/get results before sharing them, confirm note GUIDs before deletion, and be aware the stated requests-only setup is inaccurate because the script depends on Evernote/Thrift-related Python modules.

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

Tp4

High
Category
MCP Tool Poisoning
Confidence
91% confidence
Finding
The skill description claims a limited implementation ('only requests', no SDK) and a bounded feature set, but the analysis reports additional update capability and dependency on Thrift/evernote.edam interfaces. This mismatch undermines informed consent and can hide broader code execution paths or supply-chain dependencies than the user expects.

Vague Triggers

Medium
Confidence
88% confidence
Finding
Overly broad trigger phrases such as generic references to notes or Evernote can cause the skill to activate unintentionally. In a skill that can read, create, search, and delete user notes, accidental activation can expose private data or perform destructive actions without clear user intent.

Static analysis

No suspicious patterns detected.