ClawHub
Back to skill

Security audit

tech-evaluation

Security checks for vulnerabilities and agentic risk

Overview

This is a prompt-only technology/product comparison skill whose web and GitHub research behavior matches its stated purpose, with some activation and tooling cautions.

Install this if you want a structured assistant for technical or product selection. Expect it to browse the web or query GitHub for evidence; avoid giving it private repository access, API tokens, or sensitive internal product details unless you explicitly want those sources included.

SkillSpector

Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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
Findings (3)

Context-Inappropriate Capability

Medium
Confidence
92% confidence
Finding
The skill explicitly directs the agent to use `execute_code` to call the GitHub API when normal web tools are unavailable, which expands the skill from content analysis into code execution. Even if the example use is benign, introducing arbitrary code execution in a broadly triggered evaluation skill increases the chance that untrusted user input or URLs are routed into executable tooling, creating unnecessary attack surface and tool-abuse risk.

Vague Triggers

Medium
Confidence
89% confidence
Finding
The trigger set includes generic terms like “选型”, “对比”, “比较”, and “推荐”, which are common in ordinary conversation and can activate the skill outside its intended context. Over-broad activation is dangerous because it may cause the agent to invoke a tool-using workflow unexpectedly, including the research and code-execution behaviors defined later in the skill.

Vague Triggers

Medium
Confidence
95% confidence
Finding
The explicit trigger list contains several vague, high-frequency phrases and even the token “vs”, which can appear in many unrelated requests. In this skill, accidental activation is more concerning because the instructions authorize external lookup and `execute_code`-based fallback behavior, so a false trigger can lead to unintended tool use and broader exposure to adversarial input.

Static analysis

No suspicious patterns detected.