storyblok-implement-figma

Pass

Audited by Gen Agent Trust Hub on Oct 14, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill relies on several bundled shell scripts (sync-assets.sh, inspect-image.sh, preview-assets.sh, and upload-asset.sh) to execute its core workflow, including interacting with the Storyblok Management API and processing local files.
  • [EXTERNAL_DOWNLOADS]: The scripts inspect-image.sh and preview-assets.sh utilize npx to download and execute the playwright library and install the Chromium browser if they are not already available. These are established, trusted tools from well-known providers used for analyzing image properties and generating manifest previews.
  • [DYNAMIC_EXECUTION]: Several scripts use the node -e command to execute JavaScript code that is embedded directly within the shell scripts. This is used for complex tasks like image sampling and rendering that are difficult to achieve in pure shell scripting.
  • [INDIRECT_PROMPT_INJECTION]: The skill has a potential attack surface for indirect prompt injection as it processes untrusted design data from the Figma API (SKILL.md, Step 2). The workflow mitigates this risk by enforcing a structured sequence (inventory, then modeling, then code generation) and using boundary markers like design order freezing. Capability inventory includes shell scripts performing network requests and file writes (sync-assets.sh). Sanitization is implemented in the form of HTML entity escaping in the preview generator (preview-assets.sh).
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 14, 2026, 07:21 AM