Use WhoUseMyPort with OpenClaw when a blocked port needs a second pass.

This page documents the OpenClaw workflow currently shipped in the repository. The repo includes a dedicated skill, a shared install script, and a structured export flow in the current desktop source build so OpenClaw can reason about localhost conflicts from real port data instead of screenshots.

  • Workflow docs updated: 2026-03-09
  • Bundled skill: whousemyport-port-diagnostics
  • Desktop export label: Export for OpenClaw
  • Public binaries on this site remain version 1.0.0 until the next build refresh
Download SKILL.md Download desktop app

Bundled repo skill

The repository includes a first-party skill under skills/whousemyport-port-diagnostics so OpenClaw can interpret blocked-port reports consistently.

Shared install path

If OpenClaw is not using this repository as its workspace, the included installer can link the same skill into ~/.openclaw/skills.

Structured report export

The current desktop source build can export the visible port table as structured text so OpenClaw can reason over filters, PID, state, protocol, and command details.

Three ways to get the skill onto your machine

OpenClaw can discover skills from the active workspace or from a shared skills directory. This site now also hosts the raw SKILL.md file directly if you only need the prompt file and do not want to clone the repository first.

1. Use this repository as your OpenClaw workspace

If OpenClaw opens this repository directly, it can discover the bundled skill without any extra install step.

<workspace>/
skills/
  whousemyport-port-diagnostics/
    SKILL.md

2. Install the shared skill link

Use the provided command when you want OpenClaw to see the WhoUseMyPort skill outside this repository.

npm run openclaw:install-skill

# or
bash scripts/install-openclaw-skill.sh

3. Download the raw file from this website

Use this when you just want the skill file itself. Place it in a folder named whousemyport-port-diagnostics as SKILL.md.

Download SKILL.md 
<workspace>/
skills/
  whousemyport-port-diagnostics/
    SKILL.md

# or
~/.openclaw/skills/
  whousemyport-port-diagnostics/
    SKILL.md

Recommended workflow

  1. Open WhoUseMyPort and narrow the table to the blocked port or suspicious process.
  2. If your build includes Export for OpenClaw, export the current view after you apply the relevant filters.
  3. Paste the report into OpenClaw after enabling the whousemyport-port-diagnostics skill.
  4. Ask a narrow question such as which PID owns port 3000, whether a listener is safe to stop, or why a service cannot bind.
  5. Keep the command column and socket state in view before acting on OpenClaw's recommendation.

If the public binary you downloaded from this site does not yet include the export button, use the current source build or manually provide the same fields: port, protocol, state, process, PID, and command.

What the exported report contains

  • A summary header with timestamp, platform, visible row count, and language.
  • The active search term plus protocol and state filters that shaped the visible result set.
  • The visible port rows, including port, protocol, state, process name, PID, and command.

That structure matters because it lets OpenClaw reason about missing rows, filtered views, and whether the relevant socket is a long-lived LISTEN entry or just transient traffic.

Why this is better than a screenshot

  • OpenClaw can read the full command text instead of guessing from cropped UI elements.
  • Filtered state is explicit, so the agent knows whether rows are missing because of search or protocol filters.
  • The report can be copied into email support as well, which is easier to quote and compare than an image.

Plugin note

OpenClaw also supports plugin-supplied skills. The current WhoUseMyPort integration keeps the setup simpler: a repository skill plus a structured desktop export. That keeps the workflow transparent and easy to audit from this codebase.