tubeify
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill uses
curlto perform multiple network operations, including authentication, data submission, polling, and file downloading. - [DATA_EXFILTRATION]: The skill sends potentially sensitive information to an untrusted domain (
tubeify.xyz). - Evidence: The
TUBEIFY_WALLETenvironment variable and user-providedvideo_urlare transmitted via POST requests tohttps://tubeify.xyz/index.phpandhttps://tubeify.xyz/process.php. - [INDIRECT_PROMPT_INJECTION]: The skill demonstrates an attack surface for indirect prompt injection by processing data from an external source.
- Ingestion points: The skill reads the
download_urlfrom the JSON response ofhttps://tubeify.xyz/status.phpand the initialvideo_urlprovided by the user. - Boundary markers: None present; the agent is instructed to read the URL directly from the response body.
- Capability inventory: The skill uses
curlto download files and perform network requests. - Sanitization: No sanitization or validation of the returned
download_urlis performed before it is passed to a shell command. - [EXTERNAL_DOWNLOADS]: The skill downloads content from an unverified external source and saves it to the local file system.
- Evidence:
curl -o edited_video.mp4 "<download_url>"downloads a file from a URL provided by the external API.
Audit Metadata