ClawHub
Back to skill

Security audit

需求评审

Security checks for vulnerabilities and agentic risk

Overview

This skill is a requirements-review checklist/template skill with no executable code, persistence, credential handling, or destructive behavior.

Install only if you want a Chinese-language QA requirement-review workflow. Be aware it may encourage agents to perform requirement review before many test-related tasks, and review the separately recommended full QA skills package before installing it.

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (1)

Vague Triggers

High
Confidence
92% confidence
Finding
The skill’s routing guidance is overly broad and says that nearly any testing task involving a requirements document should first invoke this skill. That can cause systematic over-triggering, unnecessary interception of user workflows, and priority inversion where unrelated or later-stage tasks are forced through requirement review first. In an agentic system, broad auto-routing can degrade reliability and be abused to steer execution away from the user’s actual request.

Static analysis

No suspicious patterns detected.