ClawHub
Back to skill

Security audit

检验抽样技能

Security checks for vulnerabilities and agentic risk

Overview

This skill performs local inspection-sampling data processing with disclosed local file reads and writes, and I found no hidden exfiltration, credential use, or unsafe automatic behavior.

Install only if local storage of sampling schemes, outputs, and audit records in the workspace is acceptable. Avoid using sensitive production datasets unless you are comfortable with parsed data, sample selections, source paths, and reports being written to local JSON files.

SkillSpector

Vulnerability Patterns
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Lp3

Medium
Category
MCP Least Privilege
Confidence
94% confidence
Finding
The skill clearly instructs use of scripts that read input files and write schemes, outputs, and persistent records, but no permissions are declared. This creates a transparency and governance gap: an agent could access or modify user workspace files without an explicit capability declaration or user-facing expectation of that access.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill documents persistent storage of schemes and sampling records under the user workspace, but does not prominently warn that operational data, logs, and sample details will be written to disk. This can expose sensitive inspection data, create retention/privacy issues, and surprise users who expected transient analysis only.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The documented delete command enables destructive modification of stored schemes without any warning, confirmation, backup, or recovery guidance. If invoked accidentally or through ambiguous prompting, users could lose configuration data and auditability of sampling workflows.

Static analysis

No suspicious patterns detected.