ClawHub
Back to skill

Security audit

Ai Planner

Security checks for vulnerabilities and agentic risk

Overview

The skill appears to be a legitimate e-commerce image-planning helper, but it retains generated brief history locally without clear opt-in or retention controls.

Review before installing if you may provide confidential product launches, campaign plans, customer details, or business strategy. Use an isolated workspace for sensitive work, check or clear memory/brief_history.md after use, and prefer a version that makes history storage explicit and optional.

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (2)

Vague Triggers

Medium
Confidence
92% confidence
Finding
The natural-language trigger example is very broad (`帮我分析这个需求...`) and resembles ordinary conversation, which can cause accidental invocation in unrelated contexts. In agentic environments, ambiguous activation can route sensitive user text into this skill unexpectedly, causing misfires, unwanted processing, or chaining into downstream tools.

Ssd 3

Medium
Confidence
97% confidence
Finding
The agent persistently appends user-derived prompt content and related metadata to memory/brief_history.md in plain text without consent, minimization, access controls, or retention limits. If users include confidential product plans, campaign details, or personal information in requests, that data can be exposed to other local users, backups, logs, or later components that read the history file.

Static analysis

No suspicious patterns detected.