ClawHub
Back to skill

Security audit

百炼搜索工具-免费版

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly a web-search helper, but it grants shell/write-style authority while giving broad and inconsistent instructions about creating, deleting, saving, and caching data.

Review before installing. Use this only for non-sensitive web search or keyword research, avoid confidential or regulated queries, and be careful with the local save/cache examples because they can leave query topics and search results on disk.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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)

Description-Behavior Mismatch

Medium
Confidence
95% confidence
Finding
The skill is presented as a web-search tool, but the documentation broadens its capabilities into generic create/modify/delete/export/save operation modes that are unrelated to search. This kind of scope inflation can mislead an agent into applying the skill in unintended contexts and increase the chance of unsafe file or command operations under the skill's allowed tool set.

Intent-Code Divergence

Medium
Confidence
89% confidence
Finding
The skill claims the free version does not provide caching/history/archival features, yet later provides shell examples that cache queries and persist results to local files. This inconsistency can cause agents or users to store potentially sensitive search queries and externally sourced content on disk without clear consent, retention policy, or security guidance.

Vague Triggers

Medium
Confidence
86% confidence
Finding
The trigger condition is broad enough to activate the skill for loosely related SEO, ranking, and traffic-optimization tasks without sharply defining permitted behavior. In an agent setting, vague activation rules can cause overuse of external search and unintended transmission of user prompts or business-sensitive queries to third-party services.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The skill does not clearly warn that user queries are transmitted over the network to the Bailian/DashScope external API. This creates a privacy and data-governance risk because users or agents may submit confidential prompts, internal research topics, or proprietary keywords without informed consent.

Static analysis

No suspicious patterns detected.