Secrets Management

Security checks across malware telemetry and agentic risk

Overview

This is an instruction-only secrets-management guide; its examples touch sensitive systems but are coherent with the stated purpose.

Use this as a reference guide, not a copy-paste production script. Replace placeholder values, avoid dev/root tokens, do not log secret-derived values, pin third-party CI actions and container images, and review any command that creates, rotates, or injects secrets before running it against production.

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)

Intent-Code Divergence

Medium

Confidence: 94% confidence
Finding: The example logs `Connecting to database as $DB_USERNAME`, which exposes a secret-derived identifier in CI logs. While a username is typically less sensitive than a password or token, it still reveals account names and environment details that can aid reconnaissance or correlate with other leaked credentials.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal