Security audit

提醒

Security checks across malware telemetry and agentic risk

Overview

This reminder skill is not clearly malicious, but it can create persistent scheduled actions and external notifications without tight scoping or clear user consent.

Review this skill carefully before installing. Only use it if you are comfortable with an agent creating SkillHub cron tasks, and require explicit confirmation of the reminder text, scheduled time, destination, recurrence, and deletion behavior before any command runs. Avoid using it for sensitive reminder content unless the external channel and retention behavior are fully understood.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
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
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (6)

Intent-Code Divergence

Medium

Confidence: 92% confidence
Finding: The skill is presented as creating one-time reminders, but the FAQ expands scope to recurring reminders and task modification/deletion. This mismatch can cause an agent or user to authorize more persistent or broader actions than expected, increasing the chance of unintended scheduled tasks or lifecycle operations.

Intent-Code Divergence

Medium

Confidence: 95% confidence
Finding: The parameter section describes generic content-processing fields instead of reminder-specific inputs such as schedule time, destination, and reminder text. This ambiguity can lead an agent to infer missing parameters, mishandle user intent, or construct unsafe commands with incorrectly validated input.

Intent-Code Divergence

Low

Confidence: 88% confidence
Finding: The output schema is a generic processing-result structure rather than a reminder-task creation result containing task ID, scheduled time, destination, and persistence status. This can obscure whether a persistent task was actually created and make auditing, user confirmation, and safe rollback harder.

Intent-Code Divergence

Medium

Confidence: 93% confidence
Finding: The description mixes reminder scheduling with unrelated marketing-copy and content-writing use cases, making the skill's operational boundary unclear. Overbroad descriptions increase the risk of unintended invocation, misuse of exec-capable tooling, and user misunderstanding about what actions the skill may perform.

Vague Triggers

Medium

Confidence: 90% confidence
Finding: Overly broad and inconsistent invocation conditions make it easier for an agent to trigger this skill in contexts unrelated to reminder scheduling. Because the skill advertises exec/write capability and external task creation, accidental invocation can result in unintended persistent actions or external notifications.

Missing User Warnings

Medium

Confidence: 96% confidence
Finding: The skill does not clearly warn that it creates persistent scheduled tasks and may send reminders to external channels, despite the example showing channel delivery and task persistence. Missing this warning undermines informed consent and can lead to surprise notifications, data disclosure to third-party channels, or lingering scheduled actions after the session ends.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.