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