ClawHub
Back to skill

Security audit

smart-economics-skill

Security checks for vulnerabilities and agentic risk

Overview

The skill is an educational economics coach, but it automatically sets up recurring daily messages and can proactively steer broad everyday conversations without clear user controls.

Install only if you want a Chinese-language economics coach that may proactively add economic framing to everyday decisions and send a recurring 08:00 daily prompt. Before installing, confirm you can review, pause, modify, or remove the scheduled task, and do not rely on its debt or investment content as personalized financial advice.

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 (3)

Vague Triggers

Medium
Confidence
91% confidence
Finding
The trigger keywords are broad, common phrases such as '选哪个', '做不做', '想买', and '以后再说', which are likely to appear in ordinary conversation outside the intended domain. This can cause unsolicited skill activation and response hijacking, reducing user control and potentially overriding the assistant’s primary task with irrelevant economic reframing.

Vague Triggers

Medium
Confidence
89% confidence
Finding
The instruction to 'proactively reinterpret' user conversations whenever they involve listed scenarios lacks clear boundaries on when activation is appropriate. In an agent setting, ambiguous proactive behavior can lead to over-triggering, unwanted intervention, and context takeover, especially because many listed scenarios are common life discussions rather than narrowly scoped commands.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The document states that installing the skill will automatically create a daily 08:00 scheduled task, which is a system-changing action performed without clear user consent or an explicit warning at the point of install. Even though the action appears related to the skill’s intended function, silent persistence can surprise users, create unwanted notifications or automation, and normalize background changes without approval.

Static analysis

No suspicious patterns detected.