ClawHub
Back to skill

Security audit

BigML

Security checks across malware telemetry and agentic risk

Overview

This BigML skill is a coherent connector that discloses its read, write, and delete actions and requires confirmation before state-changing operations.

Install this only if you want Codex to operate your BigML account through OOMOL. Review payloads before approving writes, and be especially careful with delete_prediction because it permanently removes a stored BigML prediction resource.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Vague Triggers

Medium
Confidence
90% confidence
Finding
The invocation rule is excessively broad: 'Use this skill for ANY BigML request' can cause the agent to route all BigML-related tasks through a high-privilege connector, including write and destructive operations, without sufficient task-level narrowing. In context, this increases the chance of accidental execution of state-changing actions when the user intent is ambiguous or when a safer, more constrained path should be chosen.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.