skills/liangmiqwq/skills/antfu/Gen Agent Trust Hub

antfu

Pass

Audited by Gen Agent Trust Hub on May 4, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides numerous shell commands for project automation, including dependency installation (ni), script execution (nr), and linter execution (eslint --fix).
  • [EXTERNAL_DOWNLOADS]: The documentation includes instructions to fetch and execute packages from the npm registry using npx (via the nlx alias) and references reusable GitHub Actions workflows from the sxzz/workflows repository.
  • [REMOTE_CODE_EXECUTION]: Utilizes npx to download and run the fast-npm-meta utility and simple-git-hooks directly from the package registry.
  • [PROMPT_INJECTION]: Indirect prompt injection surface identified. The agent processes project source files and executes tools that could be subverted by a malicious repository configuration.
  • Ingestion points: Local project source files and configuration (package.json, tsconfig.json).
  • Boundary markers: Absent.
  • Capability inventory: Shell command execution and local filesystem access (via alias.ts script).
  • Sanitization: Absent.
Audit Metadata
Risk Level
SAFE
Analyzed
May 4, 2026, 09:03 AM
Security Audit — agent-trust-hub — antfu