Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

ecommerce-image-suite

v1.0.6

电商套图生成助手。用户明确提出需要生成电商套图、商品主图、卖点图、场景图、模特图等图片内容时触发。 支持国内平台(淘宝、京东、拼多多、抖音)与国际跨境平台(Amazon、独立站)的尺寸规范。 触发示例:「帮我生成这件T恤的电商套图」「做一套淘宝主图」「生成亚马逊listing图片」。 不应在用户仅上传图片但未明确...

1· 124·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
Capability signals
Requires OAuth token
These labels describe what authority the skill may exercise. They are separate from suspicious or malicious moderation verdicts.
VirusTotalVirusTotal
Pending
View report →
OpenClawOpenClaw
Suspicious
medium confidence
Purpose & Capability
The skill's name/description match the code: it builds prompts, calls image-generation providers, and overlays copy. Requiring a provider API key (DASHSCOPE_API_KEY) is reasonable because the skill targets multiple AIGC providers and lists DashScope/Tongyi as a primary vendor. However there are internal inconsistencies in environment variable names (see instruction_scope/environment_proportionality) which will cause confusion or misconfiguration.
!
Instruction Scope
Runtime instructions and scripts explicitly send product descriptions and (implicitly) uploaded product images to external image-generation APIs. The code accepts custom --base-url values (proxy endpoints) and will send Authorization headers or API keys to those URLs; this creates a clear exfiltration vector for both API keys and image/prompt data if a malicious proxy is supplied. The SKILL.md also references using Claude Vision for analysis (prompts provided) but there is no declared env/config for Anthropic — the skill expects the agent/platform to supply that model, which is plausible but worth noting.
Install Mechanism
No install spec; scripts are instruction-driven Python scripts. Dependencies (requests, Pillow) are required but not automatically installed. No remote arbitrary archive downloads or installers were found in the manifest.
Credentials
Requiring a provider API key is proportionate. The manifest declares DASHSCOPE_API_KEY as required and several other provider keys as optional, which fits a multi-provider generator. However, there are naming inconsistencies: generate.py/refs use ARK_API_KEY / ARK_BASE_URL for the 'doubao' provider in some places while check_providers.py lists DOUBAO_API_KEY / DOUBAO_BASE_URL (typo/inconsistent naming). This can cause failure to detect configured keys or confusion about which secret you must provide. check_providers prints a key_preview (first 8 chars) which may leak partial keys into logs/output.
Persistence & Privilege
always:false and no evidence the skill modifies other skills or system settings. It does not request elevated system presence.
What to consider before installing
This skill appears to implement the advertised e‑commerce image workflow, but review and be cautious before installing or running it: - Confirm which provider key you will supply. The skill lists DASHSCOPE_API_KEY as required (Tongyi / DashScope). Optional keys for OpenAI, Google Gemini, Stability, ARK/豆包 are also recognized — only set the keys you intend to use. - There are inconsistent environment variable names in the code/documentation (e.g., DOUBAO_API_KEY vs ARK_API_KEY). Double‑check the script you run (check_providers.py vs generate.py) to ensure the environment variable names you set are actually detected. - The generate scripts allow custom --base-url / *_BASE_URL values (proxy endpoints). If you use a proxy URL, the script will send your API key and prompt/image data to that proxy. Only use trusted proxy endpoints (prefer official provider endpoints) — untrusted proxies can harvest your API keys and uploaded images. - The skill will send product images/prompts to external services (image generation and the 'visual analysis' step). Do not supply sensitive or private images unless you accept they will be transmitted to those providers. - check_providers prints a key preview; avoid running it where its stdout could be captured by others if you care about leaking partial key data. - If you plan to use this in production, consider: (1) fixing the env var name mismatches in the code, (2) adding explicit validation/whitelisting for allowed base_url values, and (3) limiting logging of API keys/partials. If you are unsure about trusting provider endpoints or proxies, test the skill in a controlled environment with non-sensitive images and short‑lived API keys that you will rotate afterward.

Like a lobster shell, security has layers — review code before you run it.

latestvk97apwebrvje9fqbtc3bjz356x84rk2d

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

Runtime requirements

🛍️ Clawdis
EnvDASHSCOPE_API_KEY
Primary envDASHSCOPE_API_KEY

Comments