web-artifacts-builder

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEPRIVILEGE_ESCALATIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [PRIVILEGE_ESCALATION]: The script scripts/init-artifact.sh attempts to install the pnpm package manager globally using npm install -g pnpm if it is not detected. Global installations can be used to escalate privileges or modify the environment in ways that persist across sessions.
  • [COMMAND_EXECUTION]: The skill relies on executing various shell commands and CLI tools for its core functionality.
  • scripts/init-artifact.sh uses pnpm, vite, sed, tar, and node -e to bootstrap the project.
  • scripts/bundle-artifact.sh uses pnpm exec parcel and pnpm exec html-inline to build and package the artifact.
  • [EXTERNAL_DOWNLOADS]: The initialization and bundling scripts automatically install a large number of dependencies from the npm registry. While these are common frontend libraries (React, Tailwind, Radix UI), automated installation of external packages increases the attack surface for supply chain issues.
  • [INDIRECT_PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection.
  • Ingestion points: User-modified code in the project directory created by scripts/init-artifact.sh.
  • Boundary markers: None present; the agent is instructed to develop by editing generated code and then bundling it.
  • Capability inventory: The skill can execute build commands (parcel), file writes, and network operations (via package installation).
  • Sanitization: No sanitization is performed on the code before it is bundled into the final HTML artifact.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 07:18 PM