stitch-ui-react-native-components
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local shell script (
scripts/fetch-stitch.sh) to retrieve design assets. This script utilizescurlwith security-hardened flags (e.g.,--proto '=https') to download HTML and screenshots. - [EXTERNAL_DOWNLOADS]: The workflow involves downloading external HTML files and images from Google Cloud Storage based on metadata retrieved from project connectors. These downloads are essential for the conversion process.
- [DYNAMIC_EXECUTION]: The skill includes a Node.js validation tool (
scripts/validate.js) that uses the@swc/corelibrary to perform Abstract Syntax Tree (AST) analysis on generated components. This is used to ensure compliance with architectural rules and React Native best practices. - [INDIRECT_PROMPT_INJECTION]: The skill processes external HTML content, which represents a potential surface for indirect prompt injection. The risk is mitigated by explicit instructions to sanitize PII and the use of structural mapping rules (HTML to React Native primitives) that limit the influence of arbitrary embedded instructions.
- [REMOTE_CODE_EXECUTION]: Python test scripts (
tests/test_fetch.py) use thesubprocessmodule to execute the fetch script during local testing. This is a standard testing practice and does not represent a threat to the user environment.
Audit Metadata