ClawHub
Back to skill

Security audit

senior-data-scientist

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this skill suspicious because the scanners report Kubernetes and Helm deployment actions beyond its stated data science purpose, and some referenced files were unavailable for review.

Consider whether you need this skill's deployment capabilities before installing it. The scan is partial: referenced scripts and manifests were missing, so their behavior remains unassessed.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

Model training, evaluation, testing, and linting are consistent with or supportive of the declared data science workflows. However, building a service image, applying Kubernetes resources, upgrading a Helm release, and monitoring a deployed service introduce infrastructure deployment and operations capabilities not disclosed in the description. The contents of the referenced scripts and manifests are unavailable, so their additional behavior cannot be assessed.

Content

No source excerpt is available for this finding.

The skill includes Kubernetes deployment commands outside its stated statistical analysis and model-development purpose.

Medium
Category
Excessive Agency
Confidence
91% confidence
Finding

The deployment commands at L217–L218 apply Kubernetes resources and upgrade a Helm release, potentially changing running services in the active cluster context. The manifest describes experimentation, causal analysis, feature engineering, model evaluation, and MLflow tracking—not infrastructure deployment or service administration.

Content

No source excerpt is available for this finding.