brand-guide-visualizer
Pass
Audited by Gen Agent Trust Hub on Aug 24, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The generated HTML visualizer is permitted to load CSS from Google Fonts and Fontshare. These are established services for web typography and are treated as trusted sources.
- [COMMAND_EXECUTION]: The skill provides the user with platform-appropriate shell commands (such as
openon macOS orstarton Windows) to view the generated file. This is a standard utility function for local file access. - [DATA_EXPOSURE]: The skill reads project-specific brand guidelines from the local file system (
brand-guide.md) and saves the resulting artifact locally. No credentials or sensitive system files are targeted. - [INDIRECT_PROMPT_INJECTION]: The skill processes content from
brand-guide.mdand uses it to dynamically generate an HTML file containing JavaScript. - Ingestion points:
brand-guide.mdfile in the current working directory or/mnt/user-data/uploads/. - Boundary markers: The skill relies on strict 'Structural contract' rules to define the output, which acts as a behavioral constraint for the agent during generation.
- Capability inventory: The skill possesses the ability to read from the file system and write new files to the disk.
- Sanitization: There are no instructions provided to sanitize the markdown content before it is interpolated into the generated code.
Audit Metadata