The Agent Skills Directory

[PRIVILEGE_ESCALATION]: The skill utilizes sudo to bind the HTTPS proxy to privileged port 443. It also provides commands like portless trust to modify the system-wide certificate trust store and portless hosts sync to modify the /etc/hosts file.
[PERSISTENCE]: The portless service install command registers persistent startup services across multiple operating systems, including macOS launchd, Linux systemd, and Windows Task Scheduler.
[COMMAND_EXECUTION]: The core functionality relies on portless run <cmd>, which executes arbitrary shell commands provided by the user or the environment, potentially allowing the execution of dangerous payloads if input is not validated.
[EXTERNAL_DOWNLOADS]: The skill recommends installing the portless CLI tool from external registries such as npm (npm install -g portless) or Homebrew, and mentions sharing apps publicly via ngrok or Tailscale tunnels.
[INDIRECT_PROMPT_INJECTION]: The skill provides an attack surface where an agent might be directed to execute high-privilege commands or modify system configurations based on untrusted data.
Ingestion points: User-supplied commands passed to portless run in SKILL.md examples.
Boundary markers: None identified.
Capability inventory: Shell execution, sudo access, system-wide certificate store modification, /etc/hosts modification, and startup service registration.
Sanitization: None identified; the skill relies on the underlying portless CLI tool.

portless