ClawHub
Back to skill

Security audit

Morning Briefing

Security checks across malware telemetry and agentic risk

Overview

This looks like a legitimate morning briefing skill, but it deserves review because it can automatically read private work context and send it on a schedule without clear preview or consent controls.

Install only if you intentionally want an automated daily briefing that reads your work notes and calendar context and sends a summary to the stated recipient. Before enabling it, confirm the cron job is opt-in, verify exactly which files and services it can read, narrow the trigger phrases, and require preview or confirmation before sending sensitive project details.

SkillSpector

By NVIDIA
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 phrases include broad, everyday terms like '今天怎么样' and general status-report wording, which can cause unintended invocation during normal conversation. Because this skill aggregates multiple internal data sources and sends a briefing automatically, accidental activation can expose project status, logs, or schedule information without clear user intent.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The skill instructs the agent to collect data from multiple sources such as memory logs, calendars, and task systems, then send a synthesized report to another recipient without an explicit consent or disclosure step. This creates a real data-leakage risk because sensitive operational details may be shared automatically or in response to a vague request, especially in a scheduled context.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal