nvm-troubleshooting-macos

Pass

Audited by Gen Agent Trust Hub on Apr 6, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Executes shell commands in SKILL.md to perform diagnostics, such as checking environment variables ($SHELL, $PATH, $NVM_DIR) and verifying tool versions (node, nvm).
  • [DATA_EXPOSURE]: Accesses local configuration files (~/.zshrc) to identify missing or misconfigured nvm initialization logic. This involves reading sensitive system files, but usage is limited to grepping for relevant configuration patterns.
  • [EXTERNAL_DOWNLOADS]: References official documentation from the nvm-sh repository on GitHub in the examples directory for additional troubleshooting context.
  • [PROMPT_INJECTION]: The skill processes the contents of user shell configuration files during diagnostics, which presents a surface for indirect prompt injection.
  • Ingestion points: ~/.zshrc is read via cat in SKILL.md.
  • Boundary markers: Absent.
  • Capability inventory: Skill executes shell commands to modify profile files and verify system state.
  • Sanitization: None; the agent greps content directly from the file.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 6, 2026, 10:50 AM