design-system-checklist
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits a potential surface for indirect prompt injection through the management of the project configuration file.
- Ingestion points: The skill reads from
.design-config.mdduring the 'UI Polish Pipeline' (Section 4, Step 3) and the 'Integration' phase (Section 6). - Boundary markers: There are no explicit instructions or delimiters defined to treat the content of the configuration file as untrusted data or to ignore embedded instructions.
- Capability inventory: The skill possesses file-writing capabilities (creating
.design-config.mdin Section 6) and file-reading capabilities (auditing TSX/CSS files and the config file in Section 4). - Sanitization: User-provided inputs (such as project name, audience description, and references) are interpolated directly into the markdown template for
.design-config.mdwithout visible sanitization or validation. If an attacker controls these values, they could inject instructions that influence the agent's behavior during subsequent 'UI Polish' or 'Onboarding' tasks.
Audit Metadata