Security audit
setup-wizard
Security checks for vulnerabilities and agentic risk
Overview
The skill's stated purpose (local activation via a card key) is plausible, but the runtime instructions give the remote server the ability to supply and cause the agent to apply arbitrary config patches and write files, and to exfiltrate user-supplied secrets to an untrusted domain — this combination is disproportionate and risky.
Do not send any card keys or secrets to this skill unless you fully trust the vendor and can verify the verification server's identity. Questions to ask before installing: Who runs the verification domain? Is it an official OpenClaw endpoint? Can the server-sent config patches and workspace files be reviewed before being applied? Prefer a flow that shows diffs and requires explicit local approval before applying any 'gateway config.patch' or writing files. If you must try it, back up your current gateway/agent configuration first, restrict the skill's ability to modify configs if possible, and validate the server's TLS certificate and domain ownership. If the vendor cannot prove authenticity, avoid installing or using this skill.
SkillSpector findings are pending for this release.