ClawHub
Back to skill

Security audit

Openclaw Skill Observability

Security checks across malware telemetry and agentic risk

Overview

This is a narrow OpenClaw observability skill that reports recent costs and errors, but its error report may reveal sensitive local log details.

Install this only where the people invoking the skill are allowed to see local OpenClaw operational telemetry. Review outputs before sharing them, because raw log lines and failed-session summaries may include paths, stack traces, session IDs, titles, usage details, or accidentally logged secrets.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Context-Inappropriate Capability

Medium
Confidence
90% confidence
Finding
The skill reads host-level systemd journal logs via `journalctl`, which exposes potentially sensitive operational data such as internal errors, service state, paths, tokens accidentally logged, or other users' activity. Given the stated code purpose is reporting session costs and recent errors, broad host log access is more privileged than necessary and increases the blast radius if the skill is misused or compromised.

Context-Inappropriate Capability

Medium
Confidence
86% confidence
Finding
The skill uses `child_process.exec`, which grants shell execution capability inside the host environment. Even though the current commands are mostly fixed strings, this unnecessarily expands privilege and creates command-injection risk if any argument, such as `limit`, ever becomes attacker-controlled or is modified in future changes.

Missing User Warnings

Medium
Confidence
84% confidence
Finding
The skill executes external shell commands without any user-facing notice that it will access local system resources and invoke host binaries. This reduces transparency and can surprise users or operators about the level of host access the skill has, especially in environments where subprocess execution is security-sensitive.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The skill accesses system logs without disclosing that it will inspect host-level journal entries. Because logs may contain sensitive operational or user data, undisclosed collection and presentation of this information increases privacy and data-exposure risk.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.