linkpix-seedance-2-sales
Fail
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: HIGHINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides instructions to download the Node.js runtime from the official
nodejs.orgdistribution server. The procedure includes a mandatory step to verify the file's integrity using SHA-256 checksums before installation, which is a security best practice. - [DYNAMIC_EXECUTION]: To handle image size limitations, the skill instructs the agent to generate and execute a Python one-liner using the
Pillowlibrary to resize and compress local image files. This is a defined utility function but involves runtime code generation based on file paths. - [COMMAND_EXECUTION]: The skill relies on executing the vendor's CLI tool (
qhkit) for video generation and status checking. It also includes instructions for managing dependencies vianpmandpip. - [INDIRECT_PROMPT_INJECTION]:
- Ingestion points: The skill ingests untrusted data in the form of user-provided text prompts, and local image and video files (specifically
uploadedImages,uploadedVideo, anduploadedAudioparameters defined inSKILL.md). - Boundary markers: The instructions use JSON formatting for CLI tool parameters, which provides structured encapsulation but does not inherently sanitize embedded natural language instructions within the data.
- Capability inventory: The skill possesses capabilities to read and write files (for image processing), execute shell commands (
qhkit), and perform network requests (via the CLI tool and installation commands). - Sanitization: The skill includes explicit instructions for the agent to validate and compress image files if they exceed the 10MB limit before passing them to the generation tool.
Recommendations
- HIGH: Downloads and executes remote code from: https://nodejs.org/dist/v22.22.3/SHASUMS256.txt - DO NOT USE without thorough review
Audit Metadata