Use GoPlus AgentGuard safely with AI
Works out which AgentGuard tool fits your question, reports each verdict with its basis, and asks before changing any trust entry.
Use it when
- You want to check whether an agent skill or plugin is safe before installing it.
- You need to know whether one command or file write should run.
- You are reviewing which skills your local agent already trusts.
Tools it uses
skill_scanner_scanregistry_lookupregistry_listregistry_attestregistry_revokeaction_scanner_decideaction_scanner_simulate_web3
Can I safely install this skill called log-helper, and is the shell command it wants to run allowed?
Scanned log-helper and found 2 medium findings: one network call to an unlisted host and one file write outside the project. Its registry entry is not trusted yet. The shell command was decided as deny by the local rule for piping a remote script into bash. I have not changed any trust entries. Say yes if you want me to attest log-helper after you have reviewed the two findings.