e2e-test-specialist

Security checks across malware telemetry and agentic risk

Overview

This is an instruction-only end-to-end testing skill whose examples fit testing work, though users should keep its database reset examples away from production data.

Safe to install as a testing guidance skill. Before using generated or copied code, confirm BASE_URL, API tokens, and DB_* variables point only to local, staging, or disposable test resources, and review any database cleanup steps before running them.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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 (2)

Context-Inappropriate Capability

Medium
Confidence
91% confidence
Finding
The example includes helper code that directly connects to a database and performs `TRUNCATE TABLE ... CASCADE` operations for seeding and cleanup. In an agent skill context, this expands from test-authoring guidance into destructive infrastructure manipulation; if copied into a real environment or run against misconfigured targets, it can erase data well beyond an isolated test sandbox.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The markdown presents destructive database reset logic without any cautionary note, safety guard, or statement that it must only be used against disposable test environments. In a reusable skill, omission of that warning materially increases the chance that users or downstream agents apply the pattern to shared or production-like databases, causing unintended data loss.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal