storybook
Pass
Audited by Gen Agent Trust Hub on May 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: Executes shell commands to verify project dependencies, detect platform configurations (web vs native), and manage Storybook dev server processes using 'npm run'.\n- [COMMAND_EXECUTION]: Interfaces with OS-specific keychain utilities (macOS 'security', Linux 'secret-tool', and Windows DPAPI via PowerShell) to securely manage authentication credentials without using plain-text files or environment variables.\n- [COMMAND_EXECUTION]: Utilizes the Playwright CLI to perform visual rendering checks and take snapshots of components for verification purposes.\n- [COMMAND_EXECUTION]: Dynamically generates and executes a Playwright login script at runtime to handle authentication based on repository-specific configuration.\n- [PROMPT_INJECTION]: Exhibits an indirect prompt injection surface as the skill's primary workflow involves parsing local component source files to generate story templates and interaction tests.\n
- Ingestion points: Reads component files, TypeScript interfaces, and props during Phase 1 (Component inspection).\n
- Boundary markers: Absent; the skill does not use delimiters or instructions to ignore potential injections within component source code.\n
- Capability inventory: File system writes for scaffolding (templates/*) and shell command execution (SKILL.md, rules/playwright-cli.md).\n
- Sanitization: No explicit validation or escaping of component data is performed prior to interpolation into the generated story files.
Audit Metadata