Back to skill

Security audit

Rental Helper

Security checks for vulnerabilities and agentic risk

Overview

This rental-helper skill has useful rental-management features, but it needs Review because some network fetching and data-saving behavior is too broad or insecure, and some “real listing” data is actually hard-coded sample data.

Install only if you are comfortable with a Chinese-language rental tool that stores listing and viewing data locally. Avoid giving it private screenshots or contact details unless you want them saved, do not use untrusted URLs with the parser, and treat any built-in “real” sample listings as examples rather than verified market data. Review or fix the URL validation, HTTPS verification, save-confirmation behavior, and dependency pinning before using scraping features.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Findings (3)

T09 · Insecure Skill Coding Practices

Error
Location
scripts/parse_url.py:47
Finding

Unrestricted URL Fetching Enables Server-Side Request Forgery and Local Resource Access

Content
View full analysis
`. 3. `urllib.request.Request` accepts the URL without validating its scheme or destination. 4. `ur ...[truncated 1165 chars]
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
scripts/crawl_listings.py:23
Finding

Global Disabling of HTTPS Certificate Verification

Content
View full analysis
Remediation
View remediation

T08 · Insecure Dependencies

Warning
Location
scripts/crawl_interactive.py:56
Finding

Unpinned Dependencies and Runtime Browser-Driver Download

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (37)

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

代码的实际功能只覆盖声明中的一个子能力:‘计算租房预算’。它没有实现声明中大部分核心功能,如房源信息管理、推荐、解析链接、识别图片或抓取网站数据等。因此,虽然代码本身没有表现出危险的未声明能力,也未访问额外资源,但其实际行为与整体描述相比明显过窄,无法支撑所宣称的完整“租房助手”能力范围,属于描述与行为不一致。

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

该代码块的实际功能较窄,核心是一个命令行 Selenium 爬虫:启动无头 Chrome,访问贝壳或链家租房页面,滚动加载内容,提取房源标题、链接、价格,并在指定参数下保存到本地 JSON 文件。虽然这与声明中的“网站抓取房源”部分有关,但声明将技能描述为完整的租房助手,覆盖大量未在此代码中实现的功能,因此从‘描述是否准确代表该代码块实际行为’角度看,存在明显不匹配。此外,声明称可抓取贝壳/链家/58同城/安居客,而代码 argparse 仅允许贝壳和链家,其他平台并未真正支持。这属于能力范围上的实质性差异。

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
92% confidence
Finding

该代码与声明存在明显不匹配。声明描述的是功能丰富的租房助手,但本代码块只覆盖了“获取/保存房源数据”的很小一部分,且获取逻辑并非真实 API 接入,而是对特定城市区域返回预设样本数据。代码还会在启用 --save 时写入本地工作目录 JSON 文件。虽然“记录房源信息”与“网站抓取/网页解析”在高层目标上相关,但此代码并未实现那些能力,且其主要行为比声明狭窄得多,并包含“模拟真实 API”这一误导性实现差异,因此应判定为描述与实际行为不一致。

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

描述覆盖了大量高层功能,但此代码块的实际行为很有限:它只在特定条件下生成一组硬编码的成都商铺样例数据,按面积/租金筛选,打印结果,并可保存到本地 ~/.openclaw/workspace/rental-data/listings.json。虽然这与“房源记录/筛选”有少量关联,但与描述中的核心卖点——网站抓取、网页解析、图片识别、预算计算、对比、智能推荐、批量导入——均不符。代码注释虽提到‘真实数据’、APP 分享链接、Excel/CSV 导入,但未实际实现。因此描述明显夸大并与代码主要功能不一致。

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
93% confidence
Finding

该代码片段与声明中的“智能推荐房源”子功能部分相关,但只实现了一个非常有限的本地推荐脚本,而声明描述的是一个覆盖多场景、多数据来源的综合租房助手。从行为看,脚本不会记录房源、不会计算预算、不会生成对比表、不会解析网页或图片、不会抓取网站数据,也没有真正基于 location 做地理范围推荐。因此若将整个技能声明与该代码片段对照,描述明显比实际行为宽泛得多,存在实质性不一致。

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding

The skill advertises and operationalizes file read/write and network-driven behaviors such as local persistence, URL parsing, OCR input handling, and website scraping, but it does not declare any explicit tool scope or permissions boundaries. This creates an overbroad trust model where a host may permit powerful capabilities without clear least-privilege constraints or user-visible consent expectations.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The manifest description and all invocation examples are written exclusively in Chinese, with no indication that other languages are supported or that Chinese is a required locale for a region-specific deployment. Under the policy, a language constraint should either be optional for the user or explicitly justified.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill collects potentially sensitive personal and third-party data such as addresses, commuting patterns, contact details, viewing times, notes, and links, and later states that data is stored locally. Without an upfront privacy warning, retention policy, minimization guidance, or consent cues, users may disclose more information than intended and create durable local privacy risk.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The URL parsing and scraping flows involve contacting third-party websites using user-supplied links and query parameters, but the skill does not warn users that network requests will occur or that their inputs may be transmitted to external services. In a rental context, queries may reveal intended location, budget, employer area, or other personal preference data to third parties.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

Rental screenshots often contain phone numbers, chat handles, addresses, and other personal or third-party information. Processing OCR input without warning that these details may be extracted and potentially saved increases the risk of inadvertent collection, retention, and later disclosure of sensitive data.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The module docstring and usage guidance are written only in Chinese, and the CLI help strings throughout the script also assume Chinese-language interaction. For a general-purpose skill, this is a natural-language policy concern because it imposes a specific language without any user opt-in or documented region-specific justification.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The module docstring and all user-facing CLI descriptions are written only in Chinese, which imposes a specific language on users without offering an alternative or documenting that the skill is intentionally region-specific. This is a natural-language policy issue because it constrains language/locale without opt-in.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

This code file contains user-facing natural language in Chinese in the module docstring, CLI description, and argument help text, which effectively forces a specific language/locale. The policy for SQP-3 applies to all file types and allows locale constraints only when user choice or a clear justified region-specific limitation is documented, which is not present here.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
99% confidence
Finding

The script globally disables HTTPS certificate verification by replacing Python's default SSL context with an unverified one. This allows man-in-the-middle interception or spoofing of outbound requests to rental sites, so scraped content could be tampered with and malicious URLs or false listing data could be ingested and saved.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

This Python file contains natural-language documentation and runtime messaging entirely in Chinese, beginning with the module docstring and continuing through user instructions. Under the policy for language/locale constraints, this is a violation because it imposes a specific language without documenting a user choice or explicit regional justification.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The module and function docs repeatedly claim to fetch 'real' listings from third-party APIs, but the implementation only returns hard-coded sample records for a specific city/area. In a rental-assistant skill, this can mislead users into treating fabricated or stale listings, contact numbers, and URLs as trustworthy market data, which creates integrity and deception risk even without code execution or direct system compromise.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The module docstring and all user-facing strings are written only in Chinese, indicating the skill is designed to operate in a single language without offering user opt-in or locale selection. The policy specifically calls for flagging language or locale constraints when they are forced rather than optional or justified as region-specific.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The script advertises fetching or importing real listings from multiple channels, but the implementation only returns hardcoded sample entries for a narrow Chengdu area set. In a rental-assistant skill, this is dangerous because users may make housing or business decisions based on data falsely presented as current, real, or field-verified, which is a form of deceptive functionality rather than a memory-safe bug.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The docstring states the function generates 'real' Chengdu shop data based on market conditions and labels entries as '实地调研', while the code constructs static synthetic records. This mismatch increases trust in fabricated data and can mislead users into believing listings, contacts, and URLs are authentic and verified.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

This is a natural-language policy issue because the skill forces a specific language for its instructions and output, which can exclude users who do not read Chinese. The file does not offer any opt-in, alternative locale, or justification that this is a China-specific or Chinese-only tool.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The natural-language description is entirely Chinese and the OCR engines are hard-coded to Chinese-oriented language packs (chi_sim/ch_sim plus English) with no opt-in or explanation that the skill is region-specific. This can violate language/locale policy when users are not given a choice or informed of the constraint.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The script persists OCR-extracted listing data, including phone numbers in the "contact" field, to a local JSON file under the user's home directory without any consent prompt, minimization, or retention controls. In this skill's context, images may contain personal data from rental ads or private chats, so silently storing that data increases privacy risk if the machine is shared, backed up, or later accessed by other tools.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The module docstring, CLI description, help text, and runtime messages are all presented only in Chinese, which imposes a language/locale choice on users. Under the policy, this is a natural-language locale constraint unless the skill offers opt-in language selection or clearly documents that it is intentionally region-specific.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The module docstring and all CLI-facing text are written in Chinese, and the script does not provide any option to select another language or indicate that it is intentionally limited to a Chinese-speaking region or audience. This is a natural-language policy concern because it imposes a specific language without user opt-in.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The module docstring and CLI usage present this script as recommending listings for a target location such as a company address. However, the actual scoring logic only considers budget, room type, parsed transport text, decoration, and area; the args.location parameter is defined but never referenced, so recommendations are not truly location-based as claimed.

Content

No source excerpt is available for this finding.