Security audit

EVR Framework

Security checks for vulnerabilities and agentic risk

Overview

This skill should be reviewed because it appears to let an agent take real system and messaging actions from broad triggers without clear limits.

Review the skill text carefully before installing. Only use it if you intend to give the agent authority to perform real local and external actions, and require explicit confirmation for commands, scheduled jobs, file changes, and any outgoing messages.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (2)

Vague Triggers

Medium

Confidence: 93% confidence
Finding: The trigger conditions are broad enough to activate during ordinary conversation and routine task completion, increasing the chance that this skill takes control in contexts where real-world actions are not appropriate. In combination with the skill's execute-for-real guidance, accidental activation can push an agent toward filesystem, shell, scheduling, or messaging operations without a narrowly scoped user confirmation step.

Missing User Warnings

High

Confidence: 98% confidence
Finding: This skill explicitly instructs the agent to perform real side-effecting actions such as writing files, executing shell commands, registering cron jobs, and sending messages, but provides no safety boundaries, approval requirements, or environment restrictions. That creates a direct path to unauthorized system modification, data loss, command abuse, or unintended communications if the skill is invoked in the wrong context or by a prompt-injected request.

Static analysis

No suspicious patterns detected.