ClawHub
Back to skill

Security audit

Oracle Cloud Infrastructure

Security checks across malware telemetry and agentic risk

Overview

This Oracle Cloud skill is coherent, but its safety rules understate privileged and state-changing actions that could run scripts or change cloud resources without explicit confirmation.

Install only if you trust OOMOL with your OCI connection and are comfortable with an agent managing OCI resources. Require manual confirmation for all resource changes and especially for `run_instance_agent_command`, including the exact target instance and script content, even though the skill does not currently mark it that way.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

Vague Triggers

High
Confidence
95% confidence
Finding
The manifest description instructs the agent to use this skill for ANY Oracle Cloud Infrastructure request, including read, write, and destructive operations. That broad routing trigger can cause the skill to be invoked in high-risk situations without enough task-specific safety gating, especially because the skill exposes destructive infrastructure actions and remote command execution.

Missing User Warnings

High
Confidence
91% confidence
Finding
The skill lists run_instance_agent_command, which executes shell or batch scripts on target hosts with the agent service account's privileges, but the overall description and safety guidance do not prominently warn about this remote code execution capability. As a result, an agent or user may treat the skill as ordinary cloud management tooling rather than a privileged execution channel into compute instances.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.