Security audit
Security checks for vulnerabilities and agentic risk
This looks like a real product-photo retouching skill, but it under-scopes how user images and downloads are handled by remote server-provided URLs.
Install only if you are comfortable sending selected product photos and prompts to AI Hive. Use a dedicated API key, avoid sensitive unreleased product images unless the service is approved for that data, run it in an isolated environment, and review or constrain returned upload/download domains before operational use.
scripts/imagegen.py:231Unvalidated Server-Controlled Upload and Download Destinations
SKILL.md:57Unpinned Third-Party Dependency Installation
The skill invokes a local Python script, installs a package, accesses files, and sends data to a remote API, yet it declares no explicit tool/permission scope. This creates an avoidable trust gap: a caller may not realize the skill needs shell, filesystem, environment, and network capabilities, increasing the chance of unintended code execution or data exfiltration if the script or surrounding environment is modified.
The manifest description says "Use this skill for 商品精修、产品精修、电商修图、瑕疵修复、质感增强、反光控制、颜色校正、背景清理..." which mixes specific product-retouch intents with broad editing phrases like "颜色校正" and "背景清理." Without constraints or negative examples, these terms could match many ordinary image-editing requests outside the intended narrow product-photo retouch workflow.
The file is written as mandatory Chinese-language instructions for use of the skill, including phrases like "必须提供参考图片" and the full operational guidance in Chinese, but it does not offer any user opt-in or alternative language behavior. Under the policy, forcing a specific language without user choice is a natural-language locale violation unless clearly documented as region-specific and justified.
The skill is marketed and configured as a product-retouch workflow that requires reference images, but the exposed generate path still accepts an arbitrary free-form prompt and forwards arbitrary model params to a general image-generation backend. In a constrained enterprise or marketplace setting, this mismatch can let callers use the skill as a broader image-generation tool than intended, bypassing policy, product-scope, or approval assumptions tied to a 'retouch-only' skill.
This code file contains user-facing natural-language text exclusively in Chinese, including the module description and later CLI prompts, with no indication that another language can be selected. The policy specifically calls for flagging language or locale constraints when the skill forces a specific language without user opt-in.
The manifest describes a focused product-photo retouch skill driven through AI Hive editing. The separate task subcommand allows querying arbitrary generation tasks by task ID, which is a broader operational API capability and not necessary to express the stated end-user purpose of product retouching itself.
No suspicious patterns detected.