software-tutorial-studio
Warn
Audited by Gen Agent Trust Hub on May 14, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill uses shell commands like
mkdir,say,ffmpeg,ffprobe, andnpxto organize files, generate TTS audio, and render video content. This creates a surface for command injection if variables like the platform name or flow slug, derived from user-provided URLs or descriptions, are not strictly sanitized before execution.- [REMOTE_CODE_EXECUTION]: The skill dynamically generates React components (.tsx) and TypeScript data files (.ts) based on input from external URLs or screenshots. These files are then executed within the Remotion framework to produce the final video, which constitutes a dynamic code execution risk.- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection through its platform analysis feature.- Ingestion points: Arbitrary content is ingested from user-provided URLs in Step 1 (Mode A).
- Boundary markers: The instructions do not define delimiters or specific constraints to prevent the agent from following instructions embedded within the analyzed web content.
- Capability inventory: The agent possesses file-writing capabilities, shell command execution, and the ability to generate executable components.
- Sanitization: There is no mention of sanitizing or filtering the content retrieved from external sites before it influences the generated narration or code structure.- [EXTERNAL_DOWNLOADS]: The skill instructions include the use of
npx shadcn@latest add, which downloads and installs components from an external registry during the skill's operation.
Audit Metadata