v2.2.0

Beike Xiaoqu Research

ReviewClawScan verdict for this skill. Analyzed May 1, 2026, 8:05 AM.

Analysis

The skill is coherent for Beike housing research, but it uses mcp-chrome with the user's logged-in browser session and can optionally send housing requirements to external AI models, so it needs careful review before use.

GuidanceThis does not show malicious behavior, and the scripts are largely consistent with the stated Beike research purpose. Before installing, make sure you are comfortable giving it mcp-chrome access to a logged-in browser session, preferably run it in a dedicated Chrome profile or tab, and review your PAL/mcporter provider configuration before using consensus mode.

Findings (4)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

Abnormal behavior control

Checks for instructions or behavior that redirect the agent, misuse tools, execute unexpected code, cascade across systems, exploit user trust, or continue outside the intended task.

Tool Misuse and Exploitation

SeverityLowConfidenceHighStatusNote

scripts/region_discover.sh

NEW_TAB_RESP=$(mcp_call ... "create_tab" ...); ... else ... echo "⚠️  无法创建新 Tab，继续使用现有 Tab: $TAB（请勿手动切换页面）"

The script tries to create an isolated Chrome tab but will continue with an existing tab if that fails, so browser state can be affected during normal operation.

User impactThe workflow may navigate or reuse an existing browser tab while scraping Beike pages.

RecommendationRun it in a separate browser window/profile and monitor which tab is being controlled.

Agentic Supply Chain Vulnerabilities

SeverityLowConfidenceHighStatusNote

SKILL.md

需要用户已安装 mcp-chrome 插件（端口 12306）+ mcporter 已配置 pal server。

The skill relies on local MCP tooling and optional PAL configuration, while the registry requirements list no required binaries, credentials, or config paths.

User impactInstallation readiness and the true local tool permissions are not fully represented by the registry metadata.

RecommendationVerify the installed mcp-chrome and mcporter/PAL sources and configurations before running the skill.

Permission boundary

Checks whether tool use, credentials, dependencies, identity, account access, or inter-agent boundaries are broader than the stated purpose.

Identity and Privilege Abuse

SeverityMediumConfidenceHighStatusConcern

references/mcp-chrome-api.md

mcp-chrome 是一个 Chrome 浏览器插件，将 Chrome DevTools 协议（CDP）封装为 MCP server，允许 AI Agent 以用户已登录的会话操控浏览器。

The skill depends on a browser bridge that operates with the user's logged-in Chrome session, which is high-impact account/session authority and broader than a narrow Beike-only credential.

User impactUsing the skill can let the agent act through the same browser session where the user may already be logged in.

RecommendationUse a dedicated Chrome profile or dedicated Beike tab, close unrelated sensitive tabs, and only run the skill when you explicitly want browser automation.

Sensitive data protection

Checks for exposed credentials, poisoned memory or context, unclear communication boundaries, or sensitive data that could leave the user's control.

Insecure Inter-Agent Communication

SeverityMediumConfidenceHighStatusNote

scripts/consensus_analyze.py

cmd = ['mcporter', 'call', 'pal.chat', '--args', args_json, '--output', 'json']

The optional consensus workflow sends prompts containing candidate housing data and user requirements to the configured PAL chat models/providers.

User impactYour budget, home-buying preferences, and scraped candidate data may be shared with whatever model providers are configured behind PAL.

RecommendationUse consensus mode only if you trust your PAL/mcporter configuration and avoid including unnecessary personal or financial details.