onboarding-video

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests user-supplied product descriptions, brand assets, and marketing copy which are interpolated into emitted TSX code templates. This creates a potential surface for indirect prompt injection or XSS. The skill includes specific mitigations: it mandates HTML-escaping of all interpolated values and references a detailed security guide (references/agent-security.md) to prevent the agent from executing instructions hidden in data.
  • Ingestion points: User inputs collected during the intake workflow defined in SKILL.md.
  • Boundary markers: Explicit instructions to escape interpolated content and treat fetched content as data, not instructions.
  • Capability inventory: Generation of React-based TSX code for the Remotion framework.
  • Sanitization: Mandatory HTML-escaping requirement for all dynamic values.
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to install dependencies using npm install. The targeted dependencies are well-known, industry-standard frameworks such as remotion and react.
  • [COMMAND_EXECUTION]: The skill provides instructions for the user to execute shell commands (npm install, npx remotion studio) to build and verify the generated project. The instructions correctly state that the code is uncompiled and requires manual user verification.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 02:22 PM