media-use

Pass

Audited by Gen Agent Trust Hub on Nov 14, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill documentation includes a command to download and install the heygen CLI via a shell script from its official domain (static.heygen.ai). This is a documented requirement for the skill's search and retrieval functions.
  • [COMMAND_EXECUTION]: The skill executes external tools heygen and ffprobe to perform media searches and metadata extraction. These executions are performed using execFileSync with explicit argument arrays, which effectively prevents shell command injection vulnerabilities.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from natural language 'intents' and project-level design files (e.g., design.md). This data is used as search queries and metadata for the media manifest. While this represents a potential injection surface, the risk is minimal as the content is not executed or interpreted as instructions by the skill or the local environment.
  • [DATA_EXFILTRATION]: The skill requires a HEYGEN_API_KEY for authentication with the HeyGen API. This key is used by the heygen CLI to authorize requests for media assets.
  • [PERSISTENCE]: The skill implements a global cache in the user's home directory (~/.media) to store and reuse downloaded media assets across different projects, which is a standard and documented performance optimization.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 14, 2026, 08:43 AM