AI-Powered Visual Regression Testing

Pass

Audited by Gen Agent Trust Hub on Feb 28, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The script scripts/analyze_diff.py executes system commands to retrieve context for visual analysis.
  • Evidence: Uses subprocess.run to call git log for retrieving commit history over the last 7 days.
  • [EXTERNAL_DOWNLOADS]: The skill requires and recommends the installation of several well-known third-party libraries and frameworks.
  • Evidence: Recommends installing @storybook/test-runner and playwright via npm, and Pillow and numpy via pip for image processing.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it incorporates untrusted data from git logs and pull request descriptions into its AI analysis pipeline.
  • Ingestion points: scripts/analyze_diff.py ingests git commit messages; SKILL.md describes ingesting PR descriptions for Claude-based analysis.
  • Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the analysis logic.
  • Capability inventory: The analysis determines whether visual changes are approved or rejected, potentially allowing an attacker to bypass regression checks via malicious commit comments.
  • Sanitization: No sanitization or filtering of commit messages or PR descriptions is performed before they are processed by the AI.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 28, 2026, 05:49 AM