Skills
Navigate
skills/leastbit/claude_skills_zh-cn/web-artifacts-builder/Gen Agent Trust Hub

web-artifacts-builder

Pass

Audited by Gen Agent Trust Hub on Sep 29, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONPRIVILEGE_ESCALATIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The init-artifact.sh script accepts a project name as a command-line argument and interpolates it directly into a sed command without sanitization.
  • Ingestion points: The project name variable $1 in scripts/init-artifact.sh.
  • Capability inventory: File writing via cat and sed, command execution via pnpm, npm, and tar.
  • Boundary markers: None.
  • Sanitization: None.
  • Risk: A malicious project name containing shell special characters or sed delimiters could inject arbitrary HTML tags into the artifact or manipulate the sed command execution flow.
  • [PRIVILEGE_ESCALATION]: The init-artifact.sh script attempts to modify the global system state.
  • Evidence: npm install -g pnpm in scripts/init-artifact.sh. Global installation often requires administrative privileges and affects the host system beyond the project scope.
  • [COMMAND_EXECUTION]: The skill makes extensive use of shell commands and Node.js execution to scaffold and build web projects.
  • Evidence: scripts/init-artifact.sh and scripts/bundle-artifact.sh execute various system utilities including pnpm, npm, tar, sed, and node.
  • [EXTERNAL_DOWNLOADS]: The skill downloads numerous frontend development dependencies from the npm registry.
  • Evidence: Multiple pnpm install and pnpm add commands in both script files target standard web development libraries and tools.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 29, 2026, 03:36 AM