brand-kit

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes live website content to build brand kits, which creates a surface for indirect prompt injection. Attackers could embed malicious instructions in HTML comments, alt text, or metadata on targeted sites. The skill mitigates this by instructing the agent to treat all fetched content as data, report embedded instructions as findings instead of obeying them, and limit its crawling to the user-specified domain.
  • [DYNAMIC_EXECUTION]: The skill renders visual brand kits using tools like Claude's Artifacts, involving the generation of HTML/TSX from extracted website text. This introduces a cross-site scripting (XSS) risk. To address this, the skill provides mandatory guidelines in references/agent-security.md for strict HTML-escaping of all interpolated values, quoting of attribute values, and the prohibition of dangerous properties like innerHTML or javascript: URLs.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 02:23 PM