design-tokens
Pass
Audited by Gen Agent Trust Hub on Jul 6, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is designed to process content from external websites, which introduces a surface for indirect prompt injection attacks.
- Ingestion points: Untrusted data is ingested from external websites when the agent fulfills the intent to "extract design tokens from this site".
- Boundary markers: Absent; there are no instructions to use delimiters or to disregard instructions embedded within the extracted website content.
- Capability inventory: The skill allows the agent to perform local repository edits and trigger deployments through the build.host API (POST /api/projects/:uuid/deploy).
- Sanitization: No explicit validation or sanitization of the extracted website data is prescribed.
- [SAFE]: The skill includes proactive safety rules, such as prohibiting the printing of API keys or secrets and requiring explicit user confirmation before performing destructive actions like deleting projects or replacing domains.
Audit Metadata