ClawHub
Back to skill

Security audit

CodeBox QR Code

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed CodeBox QR-code API integration, with the main caution that dynamic QR codes collect scan analytics.

Before installing, understand that dynamic QR codes are trackable, cost credits, and may collect analytics about people who scan them, including device and location-derived information. Use static QR codes when tracking is unnecessary, protect the CodeBox API key, and confirm update or delete actions before applying them to important QR codes.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The skill prominently advertises tracking and analytics for QR scans, including device, location, time, and event history, but does not warn users that dynamic QR codes collect potentially sensitive visitor telemetry. This can lead users to enable tracking without informed consent, creating privacy, compliance, and trust risks, especially where personal data collection is regulated.

External Transmission

Medium
Category
Data Exfiltration
Content
env:
        - CODEBOX_API_KEY
      bins:
        - curl
    primaryEnv: CODEBOX_API_KEY
    emoji: "📱"
    homepage: https://www.codebox.club
Confidence
87% confidence
Finding
The skill is designed to transmit user-provided QR content and management actions, plus API credentials, to an external third-party service via curl. While expected for an API integration, this is still a real security and privacy boundary because user data and tracking configuration leave the local environment and are processed by an external provider.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The analytics endpoint explicitly returns device, browser, OS, time-series, and location statistics, yet the documentation presents this as a routine feature without any privacy or data-handling warning. That omission increases the chance of misuse or non-compliant collection of user telemetry because operators are not prompted to consider notice, consent, or data minimization.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
Exporting scan events can expose detailed usage records that may include timestamps, behavioral history, and potentially location-related metadata, but the skill does not warn about the sensitivity of exported data. This raises the risk of over-collection, insecure downstream storage, or sharing of visitor activity data without appropriate safeguards.

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
## Rules

- Always use `$CODEBOX_API_KEY` from environment, never ask the user for it inline.
- Default to `DYNAMIC` mode unless the user explicitly asks for static QR codes.
- When the user describes a style (e.g. "Christmas themed", "blue ocean"), use the `keywords` parameter in the generate call for automatic template matching, or call the catalog endpoint first to find a matching `templateId`.
- Dynamic QR codes consume 1 credit per generation. Static QR codes are free.
Confidence
80% confidence
Finding
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The skill advertises scan analytics, scan event export, and webhook capabilities that can expose privacy-sensitive data such as device details, browser/OS metadata, location, and timestamps, but the reference provides no guidance on consent, minimization, retention, or secure handling. In a QR-code tracking skill, this omission increases the likelihood that downstream agents or users will collect and process personal data unsafely or in violation of privacy expectations.

Static analysis

No suspicious patterns detected.