Security audit

Server Guardian

Security checks for vulnerabilities and agentic risk

Overview

This server recovery skill is purpose-aligned, but it gives agents broad production-style recovery authority with automatic restarts and log deletion that need careful review before use.

Install only if you are comfortable giving this skill operational control over an OpenClaw server. Prefer running health checks first, avoid automatic full recovery or cron recovery on production systems until reviewed, and adjust the scripts so log deletion and non-OpenClaw process termination require explicit approval or are disabled.

SkillSpector

Vulnerability Patterns

Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands

Findings (6)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 95% confidence
Finding: The skill invokes shell scripts and operational CLI commands that can inspect and modify the host, but it does not declare corresponding permissions. This creates a mismatch between the skill's effective capabilities and its stated trust boundary, increasing the risk of silent privileged execution and unsafe deployment in environments that rely on permission metadata.

Context-Inappropriate Capability

Medium

Confidence: 88% confidence
Finding: The script can terminate arbitrary high-memory processes based on generic system heuristics rather than restricting actions to OpenClaw/Gateway-owned processes. In a server-recovery skill, this broad kill capability can disrupt unrelated services or critical workloads, especially because the full recovery flow may invoke it automatically when memory usage is high.

Vague Triggers

Medium

Confidence: 89% confidence
Finding: The trigger scope includes broad phrases such as any abnormal condition, slowness, missing processes, or error logs, which can cause the skill to activate in many ambiguous situations. Because the skill can run recovery scripts and restart services, overbroad invocation criteria raise the chance of unintended operational changes from vague user input or noisy telemetry.

Missing User Warnings

Medium

Confidence: 94% confidence
Finding: The skill recommends automatically executing recovery actions that restart services, clear memory caches, compress logs, and delete logs older than 30 days, but it does not clearly warn about service disruption, data retention implications, or rollback limitations. In a production server context, these state-changing actions can destroy forensic evidence, interrupt availability, or worsen incidents if triggered prematurely.

Missing User Warnings

Medium

Confidence: 84% confidence
Finding: The full recovery flow calls log rotation automatically, which includes deletion of older compressed logs without any separate warning or confirmation. In an incident-response context, silently deleting logs can destroy forensic evidence and hinder debugging, making this more dangerous than ordinary maintenance cleanup.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The script deletes compressed logs older than 30 days without confirmation or a destructive-action warning. Because this skill is intended for outage and recovery scenarios, those logs may be needed for root-cause analysis, compliance, or security investigations, so automatic deletion creates operational and forensic risk.

Static analysis

No suspicious patterns detected.