ClawHub
Back to skill

Security audit

Daily Recipe

Security checks across malware telemetry and agentic risk

Overview

This skill is a daily recipe generator with an explicitly documented optional push scheduler, not evidence of hidden malicious behavior.

Install only if you want recipe recommendations and optional scheduled pushes. Review your OpenClaw environment's handling of __OPENCLAW_CRON_ADD__ tokens, because enabling push will create recurring morning/evening jobs for the chosen user and channel.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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 (4)

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding
The skill is presented as a simple recipe generator, but the file also documents push-notification management, scheduling behavior, and outbound delivery channels that materially expand its capabilities. This kind of description-behavior mismatch is dangerous because users and reviewers may approve or invoke the skill without understanding that it can create recurring notifications and interact with external messaging platforms.

Description-Behavior Mismatch

Medium
Confidence
91% confidence
Finding
The manifest describes recipe recommendation and HTML generation, but the documentation adds user push-notification management commands that are not declared in the primary skill metadata. Hidden or undocumented operational features reduce transparency and can bypass informed consent and policy review, especially where user IDs, schedules, and delivery settings are involved.

Context-Inappropriate Capability

Medium
Confidence
89% confidence
Finding
Introducing Telegram, Feishu, Slack, and Discord delivery expands the trust boundary far beyond recipe rendering. Outbound messaging features can be abused for spam, unsolicited notifications, or data leakage if they are embedded in a skill whose stated purpose does not justify external communications.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The push-management section enables outbound notifications without corresponding warning or disclosure in the skill description, which undermines informed user consent. In context, a recipe skill is expected to generate content on request, not silently enable recurring external communications across messaging platforms.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.