ClawHub
Back to skill

Security audit

周计划自动分配日程器

Security checks across malware telemetry and agentic risk

Overview

The skill mostly matches its planning purpose, but it asks for recurring automation that can rewrite weekly plans and send full daily plans externally with some important scope and privacy contradictions.

Review this skill before installing. It is not merely a local splitter: if configured as documented, recurring jobs can change planning files and send full daily-plan contents through Feishu/private messaging. Use it only with non-sensitive plans or a trusted internal Feishu setup, fix the write-scope and privacy wording, and consider adding backups or dry-run review before enabling the Friday and push cron jobs.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Tp4

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding
The documentation materially understates and misstates the skill's behavior: it describes weekly-to-daily splitting and Feishu push workflows, but also documents additional file-rewriting behavior that mutates current and next-week planning files. This kind of description-behavior mismatch is dangerous because users may grant file-write, cron, and network permissions without understanding that existing plans are being altered automatically, increasing the chance of unintended data corruption, silent content injection, or unauthorized disclosure via scheduled workflows.

Intent-Code Divergence

Medium
Confidence
91% confidence
Finding
The document first warns that daily plans may contain sensitive business information and that Feishu export can create legal/privacy exposure, then later says the pushed content 'does not involve privacy data externalization.' That contradiction is dangerous because it can mislead users into treating outbound transmission as harmless and encourage unsafe deployment of scheduled message forwarding containing sensitive names, locations, deadlines, and notes.

Description-Behavior Mismatch

Medium
Confidence
86% confidence
Finding
The script is described as copying the current week's plan into a next-week template, but it also writes back to the current week's file by merging checked daily-plan items into the existing weekly plan. This creates an integrity risk because running the copy operation has an additional side effect on source data, which can unexpectedly alter records and make automation behavior harder to reason about.

VirusTotal

59/59 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.