tvc-kinetic-typography

Pass

Audited by Gen Agent Trust Hub on Sep 27, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The script scripts/generate_typography.py ingests untrusted data from a user-provided JSON file (--copy-json) and interpolates the headline, slogan, and description fields directly into the prompt sent to the LLM.
  • Ingestion points: scripts/generate_typography.py reads the file specified via the --copy-json argument.
  • Boundary markers: Absent. The external strings are placed inside double quotes within the f-string prompt but are not enclosed in specific delimiters or accompanied by instructions to ignore embedded commands.
  • Capability inventory: The script has the capability to write to the file system (via the --output argument) and perform network operations (calling the Volcengine Ark API).
  • Sanitization: Minimal. The description field is truncated to 100 characters, but no escaping or structural validation is performed on the input content before prompt interpolation.
  • [EXTERNAL_DOWNLOADS]: The skill requires the volcenginesdkarkruntime Python library to interact with the Volcengine Ark API. This is a legitimate dependency for the stated purpose of generating typography plans using the Doubao model.
  • [COMMAND_EXECUTION]: The skill utilizes python3 to execute the scripts/generate_typography.py script for processing data and communicating with the LLM API.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 27, 2026, 07:57 AM