web-design-guidelines
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions direct the agent to read and prioritize project-local files such as DESIGN_SYSTEM.md, theme.ts, CLAUDE.md, and files within the docs/ directory. These files are part of the external codebase being processed, which serves as a potential ingestion point for malicious instructions designed to hijack the agent's task or bypass safety constraints.
- Ingestion points: SKILL.md instructs the agent to read local project files including DESIGN_SYSTEM.md, theme.ts, theme/index.ts, CLAUDE.md, AGENTS.md, and files under docs/**/.
- Boundary markers: The instructions lack explicit directives to treat project-local files as untrusted or to wrap their content in protective delimiters.
- Capability inventory: The skill empowers the agent to modify the project's source code (Apply Mode), run tests, and perform browser checks.
- Sanitization: No sanitization or validation of the content of these local configuration files is prescribed.
Audit Metadata