Line Check

Security checks across malware telemetry and agentic risk

Overview

This is a coherent repository-audit skill with a disclosed, conditional use of local Brigade diagnostics, and no evidence of hidden installation, persistence, exfiltration, or destructive behavior.

Install this if you want a structured repository health audit. Before using it on sensitive repositories with Brigade memory or handoff data, review what `brigade handoff doctor` and `brigade memory care scan` do in your environment, because those are local command executions even though the skill itself contains no executable code.

SkillSpector

By NVIDIA

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
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

Findings (3)

Context-Inappropriate Capability

Medium

Confidence: 88% confidence
Finding: The skill is framed as a read-only repository audit, but it explicitly instructs the agent to run external CLI commands (`brigade handoff doctor` and `brigade memory care scan`). Even if intended for diagnostics, external tools can perform network access, write files, load plugins, or execute repository-defined behavior, turning a passive audit into active code/tool execution on adversarial input.

Intent-Code Divergence

Medium

Confidence: 93% confidence
Finding: The document claims the audit is 'Read-only' and says to never modify the repo, yet later directs the agent to run commands whose behavior is not guaranteed to be side-effect free. This mismatch can cause operators or higher-level agents to trust the skill as safe for passive inspection when it may actually trigger writes, network calls, or other environmental effects.

Vague Triggers

Medium

Confidence: 80% confidence
Finding: The activation criteria are very broad ('audit a repository', 'assess project health', 'decide what to work on next'), which makes the skill likely to trigger in many common repo-assistance contexts. Because the skill also contains command-execution guidance, broad invocation increases the chance that risky behavior is introduced in situations where a user only expected a lightweight review.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal