penpot-design-to-code-review
Pass
Audited by Gen Agent Trust Hub on Jul 9, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the
execute_codetool to run localized JavaScript snippets for extracting design tokens, mapping structures, and generating the final DRIFT report. These operations are restricted to data processing and Penpot API interactions. - [DATA_EXPOSURE]: The skill reads user-specified component files and rendered Storybook content to perform visual and structural comparisons. This access is driven by user intent and restricted to the scope of the design audit.
- [EXTERNAL_DOWNLOADS]: The skill includes instructions to fetch rendered component data from Storybook URLs via a browser MCP. This is a standard functional requirement for the intended use-case of comparing designs to live implementations.
- [PROMPT_INJECTION]: As the skill ingests and parses external code and markup, it is theoretically susceptible to indirect prompt injection. However, the skill implements significant mitigations, including a mandatory phased workflow, explicit normalization of data (converting units and colors to canonical forms), and multiple mandatory user checkpoints where the agent must present findings for review before proceeding.
Audit Metadata