ClawHub
Back to skill

Security audit

Kubernetes Autoscaling

Security checks for vulnerabilities and agentic risk

Overview

This autoscaling guide needs review because its examples can interrupt running services and activate workloads earlier than promised.

Review before production use. Restrict access to the intended cluster, namespace, and workload. Start VPA in Off mode to inspect recommendations, then assess redundancy and disruption protections before enabling automatic updates. Correct and test the activation threshold and replica limits. The empty static scan does not resolve these operational concerns; the artifacts show no malicious intent.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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)

L204 promises to prevent scaling from zero until the configured threshold is met, but L207–L218 do not configure an activation threshold.

Medium
Category
Excessive Agency
Confidence
98% confidence
Finding

The example claims to prevent scaling from zero until the threshold is met, but the Prometheus scaler's threshold controls scaling targets, not activation; without activationThreshold, activation defaults to zero. Consequently, positive metric values below 100 can activate the workload, causing unintended resource consumption and potentially avoidable costs. This is particularly relevant in an autoscaling configuration skill because users may directly adopt the example as an operational safeguard; there is no evidence of malicious intent.

Content

No source excerpt is available for this finding.

The skill instructs enabling VPA Auto mode without warning that it can evict and recreate running pods.

Medium
Category
Output Handling
Confidence
97% confidence
Finding

The skill recommends VPA updateMode: Auto, including as a troubleshooting fix, without explaining that automatic resource updates can evict and recreate running pods. Following this guidance on production workloads can cause availability interruptions, particularly for single-replica applications or workloads without adequate disruption protections; the autoscaling context makes Auto mode legitimate but does not eliminate the operational risk.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.