ClawHub
Back to skill

Security audit

Fuzzy Agent Daily Driver

Security checks for vulnerabilities and agentic risk

Overview

This is an instruction-only workflow guide whose broad agent patterns are disclosed and aligned with its purpose, but users should be careful with memory files, web lookups, scheduling, and outbound messages.

Install this only if you want an agent workflow reference that may lead your agent to keep local memory notes, schedule jobs, spawn subagents, run background processes, fetch web content, and send channel messages. Keep secrets out of MEMORY.md, USER.md, TOOLS.md, and daily notes; review cron jobs and spawned sessions; and require confirmation before external messages or sensitive web searches.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (3)

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The document instructs agents to automatically create and routinely update persistent memory files, including user preferences, communications channels, location/timezone, and project history, without requiring explicit user notice or consent. This creates a privacy and data-retention risk because sensitive user-related information may be stored by default across sessions and later recalled in unintended contexts.

Missing User Warnings

Low
Confidence
84% confidence
Finding
The skill instructs use of web_search and web_fetch for research but does not tell the user that external requests may transmit queries and retrieve untrusted content. In an agent context, this can lead to unintended data disclosure in search terms or over-trust of fetched content, especially if used automatically during normal workflows.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The delegation pattern includes sending messages to other sessions and external channels without a clear warning that outbound communication may occur. In a daily-driver skill, this is more dangerous because it normalizes routine cross-channel actions, which could cause accidental disclosure, spam, or unauthorized notifications if invoked without explicit user intent.

Static analysis

No suspicious patterns detected.