ClawHub
Back to skill

Security audit

敏感词筛查-CMT

Security checks for vulnerabilities and agentic risk

Overview

This skill locally checks user-chosen documents for sensitive words and saves highlighted copies, with no evidence of hidden network access, credential use, or destructive intent.

Before installing, understand that this skill reads the documents you provide and saves highlighted versions to disk. Choose an explicit output folder and keep a backup if the output folder already contains a file with the same name, especially for contracts or compliance records.

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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 (4)

Lp3

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding
The skill clearly instructs the agent to generate modified output files, which implies file-write capability, but no permissions are declared in the skill metadata. This creates a trust and governance gap: callers may not realize the skill can write files to disk, increasing the risk of unintended document modification or data handling outside expected controls.

Vague Triggers

Medium
Confidence
83% confidence
Finding
The trigger keywords include very broad terms like '检查', '审核', and '标注' without limiting them to sensitive-word screening or document types. This can cause the skill to activate in unrelated contexts, leading to unnecessary processing of user files, unintended document modification, or confusion about what operation will be performed.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The skill description says it checks files but does not prominently warn that it creates modified copies with highlights added. Users may expect read-only analysis and unknowingly permit generation of altered documents, which is especially risky for contracts, compliance records, or other sensitive files where modified copies can be mistaken for originals.

Missing User Warnings

Medium
Confidence
81% confidence
Finding
The script writes modified document copies to disk automatically, defaulting to the desktop, without explicit confirmation and without safeguards against overwriting existing files with the same name. In a file-processing skill handling potentially sensitive documents, this can cause unintended disclosure through file placement and accidental data loss or confusion through silent replacement.

Static analysis

No suspicious patterns detected.