Security audit

生活事件追踪-免费版

Security checks for vulnerabilities and agentic risk

Overview

This skill appears to be a local personal event tracker, but its instructions contain mismatched triggers and under-disclosed network-capable inputs that users should review before installing.

Install only if you are comfortable with a local CLI skill that can run commands and store personal records on disk. Before use, clarify or remove the SEO trigger text and the callback_url/network behavior so the skill is invoked only for life-event tracking and does not send notifications unexpectedly.

SkillSpector

Vulnerability Patterns

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
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (4)

Intent-Code Divergence

Medium

Confidence: 93% confidence
Finding: The skill claims to be purely local and not use external APIs, yet it accepts a callback URL for asynchronous notification. That creates a mismatch between documented trust boundaries and actual capability, which can lead an agent or user to approve behavior that permits outbound communication and possible exfiltration of event data or metadata.

Intent-Code Divergence

Medium

Confidence: 88% confidence
Finding: The documentation says the skill does not involve external API calls, but the error-handling section discusses network connectivity, proxies, and firewalls. This inconsistency suggests undeclared network behavior or at minimum misleading documentation, which is dangerous because security decisions may be made under the false assumption that no network access occurs.

Vague Triggers

High

Confidence: 91% confidence
Finding: The trigger conditions are unrelated to the skill's actual purpose and are overly broad, making it easier for an agent to invoke this skill in inappropriate contexts. In an agent environment with exec capability, ambiguous routing increases the risk of unintended command execution, mishandling of user data, or interference with unrelated workflows.

Vague Triggers

Medium

Confidence: 84% confidence
Finding: The natural-language invocation guidance is too vague, stating that the tool will automatically execute operations based on user instructions without defining firm constraints. In combination with read/exec/glob/grep tools, this ambiguity can cause overbroad or unsafe actions if the agent maps loosely related prompts into command execution paths.

Static analysis

No suspicious patterns detected.