rocketmq-kubectl

Security checks across malware telemetry and agentic risk

Overview

This is a RocketMQ administration skill with no hidden code, but it gives direct production-capable commands that can delete topics, skip messages, or inject messages without enough safety guidance.

Install only if you intentionally want an agent to help administer RocketMQ through kubectl. Before using write operations, confirm the Kubernetes context, namespace, pod, topic, consumer group, and cluster, and avoid production use unless you have approval, backups or rollback procedures, and a clear reason for the exact command.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Behavioral ASTexec() Call, eval() Call, Dynamic Import
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (3)

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The skill documents a live `deleteTopic` operation with concrete `kubectl exec` examples but provides no warning, confirmation step, rollback guidance, or scope limitations. In an agent setting, this increases the chance of accidental destructive execution against production messaging infrastructure, causing message routing failures or service disruption.

Missing User Warnings

High

Confidence: 98% confidence
Finding: The skill presents `resetOffsetByTime ... -s -1` as a way to 'clear all backlog' without clearly stating that this skips unprocessed messages and can cause permanent data loss from the consumer's perspective. Because the skill targets direct administration of RocketMQ through `kubectl exec` in a cluster, an operator or agent could apply this to production consumer groups and silently discard pending business events.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The skill documents `sendMessage` directly to arbitrary topics without warning that even 'test' messages can enter live pipelines, trigger consumers, mutate downstream systems, or pollute analytics. In a production RocketMQ context, this is more dangerous because the examples use real cluster access via `kubectl exec` rather than an isolated sandbox.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal