react-components
Pass
Audited by Gen Agent Trust Hub on Apr 12, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The
scripts/fetch-stitch.shscript downloads design HTML content from external URLs using Python'surllib.requestlibrary. While the URLs are provided dynamically at runtime, the skill instructions provide guidance on scoping this access to Stitch-provided assets. - [COMMAND_EXECUTION]: The
SKILL.mdinstructions direct the agent to execute an external utility script (python3 utilities/skill-builder/scripts/record_skill_feedback.py) to persist usage and performance metrics. This script is located outside the skill's own directory structure. - [PROMPT_INJECTION]: The skill processes untrusted HTML data retrieved from external Stitch URLs, creating an indirect injection surface. * Ingestion point:
scripts/fetch-stitch.shdownloads HTML content from design URLs. * Boundary markers:SKILL.mdincludes explicit instructions for the agent to treat download URLs as untrusted input and limit network scope. * Capability inventory: The skill utilizesRead,Write, andBashtools to generate and validate code files. * Sanitization: No programmatic sanitization or filtering is performed on the downloaded HTML content prior to structural extraction.
Audit Metadata