ClawHub
Back to skill

Security audit

Datadog

Security checks across malware telemetry and agentic risk

Overview

This Datadog skill is not malicious, but it is review-worthy because it advertises inspection while also documenting tools that can change or delete production observability resources.

Install only if you intend to grant write-capable Datadog administration through ClawLink, not just read-only inspection. Use a least-privilege Datadog connection where possible and require explicit user confirmation before muting alerts, changing monitors or dashboards, replacing host tags, creating webhooks, or deleting resources.

SkillSpector

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

Description-Behavior Mismatch

Medium
Confidence
94% confidence
Finding
The manifest and top-level description frame the skill as an inspection-only Datadog tool, but the documented toolset includes state-changing and destructive actions such as deleting monitors and dashboards, muting alerts, and creating webhooks. This mismatch can mislead users or higher-level policy systems into granting trust or autonomy inappropriate for the actual capabilities, increasing the chance of unauthorized or unsafe changes in a production observability environment.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.