skills/johnsonmao/skills/pnpm/Gen Agent Trust Hub

pnpm

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • Indirect Prompt Injection (SAFE): The skill documentation advises agents to read project configuration files like package.json, .npmrc, and pnpm-workspace.yaml. While these files represent an ingestion point for potentially untrusted data (a surface for indirect prompt injection), this behavior is a fundamental requirement for the documented tool and no instructions suggest bypassing safety filters.
  • Command Execution (SAFE): The skill documents powerful features like pnpm dlx, pnpm exec, and .pnpmfile.cjs hooks. These tools are presented for legitimate development purposes, and the skill does not include any malicious scripts that utilize these capabilities.
  • Credentials (SAFE): Security-sensitive configurations, such as registry authentication in .npmrc files, are documented using environment variable placeholders (e.g., ${NPM_TOKEN}) rather than hardcoded secrets.
  • Unverifiable Dependencies (SAFE): The skill references official GitHub Actions and standard npm packages (e.g., lodash, express). It does not specify any unknown or risky remote code sources for execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:36 PM