OpenClaw Gateway Manager

What to consider before installing or letting the agent run this skill: - Confirm source: the registry metadata shows no homepage and 'Source: unknown' while documentation references a GitHub repo. Verify the repository and author (https://github.com/seastaradmin/openclaw-gateway-manager) before cloning or executing scripts. - Review scripts first: the skill is instruction-only but includes many shell scripts that will read/write $HOME, create LaunchAgents and use rm -rf for deletes. Inspect gateway-delete.sh, gateway-create.sh and the plist template in gateway-create.sh to ensure they do what you expect. - Backup important data: even with triple-confirmation and automatic backups, deletion uses rm -rf. Make manual backups before testing on production instances. - Test in a safe environment: run check-dependencies.sh and exercise create/delete operations in a local or isolated test account before use on primary systems. - Verify openclaw and node binaries: the LaunchAgent runs node to exec openclaw. Ensure the 'openclaw' binary/package you have is trustworthy; otherwise the launched process could contact external services. - Check cross-platform claims: although metadata claims multi-platform support, the scripts are primarily macOS-focused. Don't expect full Linux/Windows parity without manual review/adjustment. - Consider manual activation: instead of granting autonomous agent invocation that could create persistent LaunchAgents, run the scripts yourself the first time or require explicit user confirmation before performing operations that install services or delete data. If you want, I can summarize the exact lines in the scripts that create or delete files and the LaunchAgent plist so you can inspect them quickly.