NYC

artifacts-builder

Warn

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS] (MEDIUM): The skill performs extensive installation of external dependencies from the npm registry.
  • scripts/init-artifact.sh installs pnpm globally and over 50 frontend development packages (React, Tailwind, Radix UI, etc.).
  • scripts/bundle-artifact.sh installs parcel, html-inline, and other bundling tools at runtime.
  • [COMMAND_EXECUTION] (MEDIUM): Complex shell scripts execute multiple system-level commands.
  • scripts/init-artifact.sh performs global software installation (npm install -g pnpm) and extracts a local archive (shadcn-components.tar.gz).
  • scripts/bundle-artifact.sh executes build processes and file system modifications.
  • [DYNAMIC_EXECUTION] (MEDIUM): The skill uses dynamic code execution to modify configuration files.
  • scripts/init-artifact.sh uses node -e to execute JavaScript strings for programmatically updating tsconfig.json and tsconfig.app.json.
  • [INDIRECT_PROMPT_INJECTION] (LOW): The skill has a vulnerability surface for indirect injection since it processes user-provided content into executable web artifacts.
  • Ingestion points: User instructions described in SKILL.md are transformed into React code.
  • Boundary markers: None are implemented in the provided scripts to distinguish instructions from data.
  • Capability inventory: File system access, network downloads (via pnpm), and subprocess execution.
  • Sanitization: No sanitization or validation of the generated artifact content is visible in the provided infrastructure scripts.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 17, 2026, 06:02 PM