ClawHub
Back to skill

Security audit

BigFood · 冰箱管家（食材识别·智能食谱·菜谱·智能菜谱）| Fridge to Table & Recipe AI

Security checks across malware telemetry and agentic risk

Overview

This is a prompt-only cooking and fridge-assistant skill with broad triggers, but no hidden execution, external backend, credential use, or persistent access.

Before installing, understand that this skill may activate for general Chinese cooking or ingredient phrases. It is otherwise limited to user-directed food image analysis and recipe suggestions, with no executable code, backend calls, credential handling, or cross-session persistence shown in the artifact.

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 (6)

Vague Triggers

Medium
Confidence
93% confidence
Finding
The trigger list contains very broad everyday phrases such as '冰箱', '菜谱', '做饭', and '食材', which can cause the skill to activate in many unrelated conversations. This increases the chance of unintended context capture and inappropriate interception of user requests meant for other skills or the base assistant.

Vague Triggers

Medium
Confidence
90% confidence
Finding
The activation rules use open-ended examples like '等' and broad conditions that allow the skill to trigger without clear boundaries. Ambiguous activation logic can make the agent invoke this skill when the user's intent is only loosely related, leading to prompt-space interference and degraded routing safety.

Overly Broad Trigger

Low
Category
Trigger Abuse
Confidence
88% confidence
Finding
The single-word trigger '冰箱' is so short and generic that it may activate on many unrelated mentions of refrigerators. While not directly exploitable as code execution, it increases accidental invocation and can interfere with correct agent routing.

Overly Broad Trigger

Low
Category
Trigger Abuse
Confidence
87% confidence
Finding
The trigger '菜谱' is a highly generic term that can occur in many benign cooking discussions not intended for this specific skill. Such broad matching can cause the skill to hijack requests that should be handled more generally or by other cooking-related flows.

Overly Broad Trigger

Low
Category
Trigger Abuse
Confidence
87% confidence
Finding
The trigger '做饭' is an everyday phrase with many possible meanings and weak specificity to this skill. This creates unnecessary activation surface and makes misrouting more likely in ordinary conversation.

Overly Broad Trigger

Low
Category
Trigger Abuse
Confidence
87% confidence
Finding
The trigger '食材' is too broad because it may appear in shopping, nutrition, restaurant, or unrelated cooking contexts. This broadness increases accidental skill invocation and unnecessary exposure of the skill's instructions in unrelated requests.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.