stitch::react-native

Pass

Audited by Gen Agent Trust Hub on Jul 21, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads HTML and image assets from URLs provided by the Stitch design platform. It uses a specialized shell script (scripts/fetch-stitch.sh) to handle these downloads, which is a standard procedure for retrieving design metadata.
  • [COMMAND_EXECUTION]: Executes local shell commands and scripts for developer workflows, including fetching assets, installing dependencies via npm install, running TypeScript compilation checks, and executing code validation scripts.
  • [EXTERNAL_DOWNLOADS]: Declares a dependency on @swc/core, a well-known and trusted library for parsing and analyzing TypeScript and JSX code, used here for automated architectural validation.
  • [PROMPT_INJECTION]: Indirect prompt injection vulnerability surface is present because the skill parses content from external HTML files to extract theme tokens.
  • Ingestion points: HTML files downloaded to .stitch/designs/{page}.html from external design URLs.
  • Boundary markers: Absent from instructions during the extraction of Tailwind tokens from HTML.
  • Capability inventory: Filesystem write access, execution of local scripts via Bash, and network access via web_fetch.
  • Sanitization: None detected in the extraction phase, although the skill uses a dedicated AST-based validation script to check the generated output components.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 21, 2026, 10:25 PM