video-report
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input (a reported video URL) and instructs the agent to write it directly into a source file (
packages/example/src/NewVideo.tsx) as a component property. - Ingestion points: User-reported video URL specified in
SKILL.md. - Boundary markers: Absent; there are no instructions to validate or encapsulate the input URL.
- Capability inventory: Modifies local file system structure and executes shell commands via
bunx remotion render. - Sanitization: Absent; the URL is placed directly into the file without validation or escaping, which could allow an attacker to inject malicious characters or code breakout sequences.
- [COMMAND_EXECUTION]: The skill executes
bunx remotion render NewVideo --log=verboseinside thepackages/exampledirectory. While Remotion is a legitimate tool corresponding to the author profile, executing rendering commands on files that incorporate unsanitized user inputs creates a vector for code execution if the input manipulates the file structure.
Audit Metadata