ClawHub
Back to skill

Security audit

COSER模卡生成器cosercard-skill

Security checks for vulnerabilities and agentic risk

Overview

This is a local cosplay/model card generator with a disclosed local history feature, but users should know it keeps some generation metadata on disk.

Install only if you are comfortable using a local photo-card generator that keeps a learning_data.json history file in the skill directory. Avoid entering contact details, exact location, body measurements, or social handles unless you intend them to appear on the generated card, and delete learning_data.json if you do not want local usage history retained.

SkillSpector

Vulnerability Patterns
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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 (4)

Tp4

High
Category
MCP Tool Poisoning
Confidence
85% confidence
Finding
The declared description focuses on card generation, but the skill also persists user preference/history data and supports auxiliary behaviors like demo generation and recommendation from past usage. That mismatch matters because users may provide sensitive profile and body-measurement data without realizing it will be retained locally and reused later, which creates privacy and consent risk.

Description-Behavior Mismatch

Medium
Confidence
95% confidence
Finding
The skill writes user/profile metadata and generation history to ../learning_data.json on every successful card generation, which exceeds the advertised card-layout/export behavior. Because the stored fields include personal identifiers and profile attributes, this creates an undisclosed persistence channel for sensitive user data and increases privacy and data-retention risk.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill requests highly sensitive personal information such as body measurements, contact details, social handles, and location, then states that it records configuration history to a local learning_data.json file. Without a clear warning, consent flow, minimization, or retention policy, this can expose private profile data to other local users, backups, logs, or accidental sharing.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The script collects and persists sensitive personal data such as name, body measurements, contact details, location, social accounts, and generation history without explicit warning or confirmation. In a card-generation skill, that context makes the issue more serious because the data is highly personal and could remain on disk unexpectedly, exposing users to privacy harm if the host machine or workspace is shared or later inspected.

Static analysis

No suspicious patterns detected.