byted-bp-seedance-viral-creative-rewrite-skill

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the subprocess module to manage internal execution flows and to run the FFmpeg utility for extracting video frames. These operations are implemented using list-based arguments rather than shell strings, adhering to security best practices for preventing command injection.
  • [EXTERNAL_DOWNLOADS]: The skill includes a dedicated runtime preparation script (scripts/ensure_runtime.py) that creates a local virtual environment and installs necessary dependencies from a requirements.txt file. This is a standard procedure for managing a skill's execution environment.
  • [INDIRECT_PROMPT_INJECTION]: The skill uses the host agent to analyze user-provided template videos and images. While the ingestion of external media content is a known vector for indirect prompt injection, the skill mitigates this by using structured JSON serialization for prompt construction and applying specific sanitization rules for creative template transfer.
  • [DATA_EXFILTRATION]: The skill transmits prompt parameters and product images to the BytePlus API at ark.ap-southeast.bytepluses.com. This transmission is the core intended function of the skill and targets the official service infrastructure of the recognized vendor.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 08:49 AM