penpot-uiux-design
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill follows security best practices by defining a clear 'SandBase boundary,' instructing the agent to keep core design work local and only use external tools for necessary generative tasks. It mandates user approval for all external calls, including displaying unit prices and cost estimates, which prevents unauthorized or hidden usage of external APIs.
- [INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it processes external files and tool outputs.
- Ingestion points: The skill inspects local Penpot design files, runtime configurations, and outputs from media generation tools (
sandbase_run). - Boundary markers: The instructions explicitly require the agent to 'Preserve the user's chosen stack, source material, and authorization boundaries.'
- Capability inventory: The skill can read local files and execute specific vendor tools (
sandbase_run,sandbase_discover) for media processing and discovery. - Sanitization: While explicit string sanitization is not detailed, the skill mitigates risk through manual user confirmation gates and strict schema-defined argument requirements for all tool executions.
Audit Metadata