Security audit
Security checks for vulnerabilities and agentic risk
This transcription skill needs review because video links can redirect it to private network data that it then uploads to its provider.
Use trusted links and media you are comfortable uploading to a paid provider. For link processing, isolate the skill from private network services until redirect restrictions are added. Its single-host network claim is inaccurate, and server deletion and no-training promises are not established by the bundled code.
scripts/utils/download.jsUnchecked redirects can disclose internal HTTP responses to the cloud service
Sensitive responses from reachable, unauthenticated internal endpoints can be transmitted to the transcription provider. The attacker need not receive the response directly for this unauthorized external disclosure to occur. Exploitation depends on such an endpoint being reachable.
Token validation can be a normal supporting implementation detail, and this utility's lack of video-processing code is not itself a mismatch. However, the promotional acquisition messages and claimed third-party token requirement introduce an undeclared commercial-access flow beyond the described video-to-text functionality. This chunk does not make network requests, verify tokens remotely, or establish that the entire skill actually stops.
Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
4. 同时传入文件路径与任务ID，优先执行 `--id`，忽略 `--file`
5. 无自定义 prompt 时，默认完整转录视频全部文字
| 用户自然语言指令 | 自动生成命令 |
| -------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------- |
| 视频提取 https://example.com/video.mp4 中的文字 | `node scripts/video2text/index.js --file "https://example.com/video.mp4"` |
| 把本地 /path/to/your/video.mp4 改成小红书风格的文案 | `node scripts/video2text/index.js --file "/path/to/your/video.mp4" --prompt "改写成小红书风格的文案"` |
Without declared permissions the skill's intent is opaque and cannot be validated.
Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
当用户用自然语言下达指令时，按以下映射生成命令，保证识别与执行一致：
| 用户自然语言指令 | 生成的命令 |
| ---------------------------------------------------- | ----------------------------------------------------------------------------------------------------------- |
| 提取 https://example.com/video.mp4 里的文字 | `node scripts/video2text/index.js --file "https://example.com/video.mp4"` |
| 总结这个视频的核心观点 https://example.com/video.mp4 | `node scripts/video2text/index.js --file "https://example.com/video.mp4" --prompt "总结这个视频的核心观点"` |
The downloader defaults removeOnStop and removeOnFail to true (L0044–L0045), deleting the destination on stop here and on failure at L1053. No comment, documentation, confirmation, or user-facing message explains this deletion; generic stop/error events do not disclose it. This can also delete an existing file supplied to resumeFromFile (L0386–L0388), rather than only a newly created download.