Live Task Pulse

• Outcome: Review
• Latest audit: May 29, 2026 at 9:23 PM UTC
• Version:
  1.1.0

Overview

This is a mostly transparent progress-tracking skill, but its local CLI can be misused to read or modify JSON files outside its intended task folder.

Review before installing. Use it only if you want automatic progress tracking and chat notifications, and avoid putting secrets, private URLs, customer data, credentials, or sensitive error text in task names or results. Before using it in a shared or sensitive workspace, patch task ID validation so reads and writes cannot leave the task directory, or restrict use to task IDs generated by the create command.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands

Findings (5)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 90% confidence
• Finding: The skill describes and instructs use of filesystem persistence and a Python CLI, implying file read/write and likely environment access, but it declares no permissions. Undeclared capabilities undermine least-privilege controls and prevent users or the platform from understanding what resources the skill may touch, which is especially risky because it stores task state on disk and encourages automatic activation.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 80% confidence
• Finding: The skill claims mandatory live push notifications, automatic activation, status-query triggers, and auto-cleanup, but the file only contains procedural instructions and examples rather than implemented safeguards or integration logic. This mismatch can cause operators to assume monitoring, notifications, and cleanup are happening when they are not, leading to silent failures, stale data retention, and misleading trust in the skill's behavior.

Vague Triggers

High

• Confidence: 84% confidence
• Finding: Trigger phrases like 'what's running', 'task status', and '任务进度' are common conversational language and may collide with ordinary user queries unrelated to this skill. Because the skill is designed to auto-send messages and reveal stored task state, false triggering can leak information about ongoing tasks or create confusing unsolicited updates.

Vague Triggers

Medium

• Confidence: 84% confidence
• Finding: Trigger phrases like 'what's running', 'task status', and '任务进度' are common conversational language and may collide with ordinary user queries unrelated to this skill. Because the skill is designed to auto-send messages and reveal stored task state, false triggering can leak information about ongoing tasks or create confusing unsolicited updates.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The skill prominently advertises automatic live push notifications and JSON persistence but does not present a clear user warning or consent mechanism about storing task data on disk and sending updates automatically. In a progress-tracking skill, stored step names, status text, URLs, and error messages can easily contain sensitive operational or personal data, so hidden persistence and push behavior materially increases privacy and data-handling risk.

VirusTotal

63/63 vendors flagged this skill as clean.