ui-pattern
Pass
Audited by Gen Agent Trust Hub on Sep 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill incorporates external content through the
$ARGUMENTSand$0placeholders into the agent context without sufficient delimiters or defensive instructions. - Ingestion points: User-provided data enters the prompt via the
$ARGUMENTS(Description) and$0(Pattern type) placeholders inSKILL.md. - Boundary markers: The instructions do not employ boundary markers, such as XML tags or explicit "ignore embedded instructions" delimiters, to isolate user input from the skill's logic.
- Capability inventory: The skill instructs the agent to read local project components (
components/ui/,components/patterns/) and generate new executable UI code. - Sanitization: There is no evidence of input validation, filtering, or escaping for the content provided in
$ARGUMENTS.
Audit Metadata