Security audit
Security checks for vulnerabilities and agentic risk
The skill is a coherent guide for a digital-human video web service, but it needs review because it handles face and voice media and includes API-key guidance without enough privacy, retention, or secret-handling detail.
Review the provider's privacy and retention terms before installing or using this skill. Only upload images, videos, and voice samples when you have clear consent from the person depicted or recorded, avoid public figures or third-party media, and keep any API key out of chats, logs, screenshots, and shared scripts.
The README instructs users to upload portrait photos/videos and even clone a personal voice, but it does not clearly warn about privacy, consent, retention, or downstream data handling. In this context, users may submit highly sensitive biometric and identity data without understanding how it is stored, processed, shared, or deleted, creating meaningful privacy and abuse risk.
The skill instructs users to upload highly sensitive personal data including face images, videos, voice recordings, and cloned voice samples, but it does not provide a clear privacy warning explaining that this data is transmitted to and processed by a third-party web service and may carry biometric/privacy risks. Because the content explicitly promotes voice cloning and digital-human generation, users may unknowingly expose biometric identifiers without informed consent or retention/transmission awareness.
The document hardcodes Chinese as the basis for core functionality, stating '中文语速实测约 5.6 字/秒' and using Chinese character counts as the default timing model, without indicating support for other languages or giving users a choice. Under the policy, language/locale constraints should be optional or explicitly justified when they are imposed.
The document expands a browser-oriented user guide into direct API and scripting instructions, which materially lowers the barrier to automated use of a paid third-party service and contradicts the stated product framing. In a skill distributed as an operational guide, this can cause users or agents to transmit data and use credentials in ways they did not expect, increasing the risk of misuse and unauthorized automation.
The file explicitly targets people who want to write scripts and provides authentication and API usage details despite claiming to only document what the web app uses. That scope creep makes the skill more dangerous because it encourages repurposing a simple browser workflow into direct scripted access without corresponding safety controls, warnings, or trust boundaries.
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
| 项 | 值 |
|---|---|
| 根地址 | `https://api.a7w.cn/api/v1` |
| 鉴权 | 请求头 `Authorization: Bearer <你的 Key>` |
| 取 Key | 到 https://api.a7w.cn/ 注册创建（1 元 = 100 点） |
| 计费单位 | 点（`rate: 100` ⇒ 1 元 = 100 点） |
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
| 项 | 值 |
|---|---|
| 根地址 | `https://api.a7w.cn/api/v1` |
| 鉴权 | 请求头 `Authorization: Bearer <你的 Key>` |
| 取 Key | 到 https://api.a7w.cn/ 注册创建（1 元 = 100 点） |
| 计费单位 | 点（`rate: 100` ⇒ 1 元 = 100 点） |
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
| 项 | 值 |
|---|---|
| 根地址 | `https://api.a7w.cn/api/v1` |
| 鉴权 | 请求头 `Authorization: Bearer <你的 Key>` |
| 取 Key | 到 https://api.a7w.cn/ 注册创建（1 元 = 100 点） |
| 计费单位 | 点（`rate: 100` ⇒ 1 元 = 100 点） |
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
| 项 | 值 |
|---|---|
| 根地址 | `https://api.a7w.cn/api/v1` |
| 鉴权 | 请求头 `Authorization: Bearer <你的 Key>` |
| 取 Key | 到 https://api.a7w.cn/ 注册创建（1 元 = 100 点） |
| 计费单位 | 点（`rate: 100` ⇒ 1 元 = 100 点） |
The document instructs use of a Bearer API key but does not warn against exposing that secret in logs, screenshots, shared prompts, client-side code, or browser tools. Because Bearer tokens are sufficient for account access and paid API consumption, omission of basic secret-handling guidance can directly lead to credential leakage and unauthorized charges or data access.
The documentation is fully Chinese-language and does not indicate that the skill is Chinese-only by design, nor does it offer any language or locale choice. Under the policy, forcing a specific language without user opt-in can be a natural-language policy violation unless the locale restriction is explicitly documented and justified.
No suspicious patterns detected.