smart-port-allocation

Fail

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • EXTERNAL_DOWNLOADS (HIGH): The skill instructs the agent to use npx -y @lionad/port-key, which fetches and executes a package from the public npm registry at runtime. The package @lionad/port-key is not from a trusted source (e.g., Vercel, Anthropic, Google), posing a risk of supply chain attack if the package is compromised.
  • REMOTE_CODE_EXECUTION (HIGH): Using npx to execute code that is not locally present or version-pinned allows for arbitrary code execution within the agent's environment.
  • COMMAND_EXECUTION (HIGH): The skill performs shell-level command execution to generate ports and check for conflicts (e.g., lsof, netstat, npx).
  • INDIRECT_PROMPT_INJECTION (HIGH): The skill is vulnerable to command injection through the project name input.
  • Ingestion points: Project names provided by the user (as seen in SKILL.md and references/port-key.md).
  • Boundary markers: Absent. The skill instructions show the project name being passed directly as a shell argument: npx -y @lionad/port-key "<project-name>".
  • Capability inventory: Execution of npx, lsof, netstat, and modification of ~/.port-key/config.json.
  • Sanitization: Absent. While the instructions suggest the agent "extract initials," there is no hard constraint preventing a malicious project name like my-project"; touch /tmp/pwned; " from being executed in the shell.
  • DATA_EXPOSURE (LOW): The skill accesses the user's home directory (~/.port-key/config.json) to manage tool settings. While not inherently malicious, it establishes a pattern of local file modification based on external inputs.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 16, 2026, 12:59 PM