chromatic-diff

Pass

Audited by Gen Agent Trust Hub on Oct 11, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Communicates with Chromatic's official GraphQL API and published Storybook hosts (chromatic.com) to retrieve build metadata and render stories. These communications target a well-known service associated with the skill's purpose.
  • [COMMAND_EXECUTION]: Executes local TypeScript scripts (scripts/resolve.ts, scripts/proxy.ts, scripts/preflight.ts) using the bun runtime. This is standard behavior for CLI-driven developer tools.
  • [INDIRECT_PROMPT_INJECTION]: The skill has a potential surface for indirect prompt injection. Ingestion points: Chromatic API responses (scripts/resolve.ts) and rendered Storybook DOM content (SKILL.md Step 4). Boundary markers: Absent. Capability inventory: Network access via fetch, local file system writes to the project directory, and browser script execution via evaluate_script. Sanitization: The ingested content is used for structural and style comparison and is not processed as instructions for the agent.
  • [DYNAMIC_EXECUTION]: Employs evaluate_script in a browser session to dynamically extract DOM structure and computed style data from external Storybook pages for diffing.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 11, 2026, 10:15 PM