kb-design-system
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to discover and read external design token files within the project environment to perform accessibility audits. This creates a surface for indirect prompt injection if the audited files contain instructions intended to influence the agent.
- Ingestion points: Local files identified by the discovery commands in SKILL.md, including tokens.json, variables.css, theme.ts, and tailwind.config.js.
- Boundary markers: Absent. The skill does not provide delimiters or instructions for the agent to ignore potential commands embedded in the token data.
- Capability inventory: The skill provides shell commands for file discovery (find, Get-ChildItem) and recommends installing external testing packages via NPM.
- Sanitization: Absent. There are no instructions for sanitizing or validating the content of the token files before processing.
- [EXTERNAL_DOWNLOADS]: The skill recommends installing development dependencies from the NPM registry.
- Evidence: SKILL.md includes commands to install @storybook/addon-a11y and @storybook/test-runner.
- [COMMAND_EXECUTION]: The skill provides shell commands to locate specific files in the local filesystem.
- Evidence: SKILL.md contains a find command for Unix-like systems and a Get-ChildItem command for PowerShell to search for design token files while excluding standard build directories.
Audit Metadata