guaikei·小红书看详情找洞察

• Outcome: Review
• Latest audit: Aug 15, 2026 · 3:39 AM UTC
• Version: 1.0.0

Overview

This skill is a disclosed Xiaohongshu public-data tool that uses guaikei.com and saves local result logs, with no hidden destructive or credential-stealing behavior found.

Install only if you are comfortable sending Xiaohongshu keywords, links, and related request data to guaikei.com and having full successful results saved locally in logs. Avoid submitting private, login-only, or sensitive URLs, and periodically delete logs if retained public-data results are not needed.

SkillSpector

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (13)

Description-Behavior Mismatch

Medium

• Confidence: 94% confidence
• Finding: The README describes materially broader capabilities than the manifest, including keyword search, large-scale data mining, competitor monitoring, trend prediction, KOL screening, and batch export. This is dangerous because reviewers and users may authorize or invoke the skill for a narrow link-detail use case while the documentation steers use toward surveillance-style collection and broader scraping workflows that were not transparently declared.

Description-Behavior Mismatch

Low

• Confidence: 89% confidence
• Finding: The README states that all task results are automatically saved to a local logs/ directory, but this persistence behavior is not disclosed in the manifest description. Undisclosed storage increases data handling risk because users may assume transient processing while URLs, content, and analysis outputs are retained on disk and potentially exposed to other local users or later exfiltration.

Description-Behavior Mismatch

Medium

• Confidence: 90% confidence
• Finding: The changelog documents capabilities such as keyword search and creator-work monitoring that materially exceed the stated skill metadata focused on Xiaohongshu link routing and content understanding. This mismatch can cause the agent or reviewers to under-scope permissions, data handling, and abuse cases, enabling unintended use of broader scraping or monitoring functions without clear disclosure.

Description-Behavior Mismatch

Medium

• Confidence: 93% confidence
• Finding: The documentation materially expands the skill from link-based Xiaohongshu routing into keyword search, competitor analysis, trend tracking, and KOL monitoring. That mismatch can cause an orchestrator or downstream agent to invoke capabilities the manifest does not declare, weakening least-privilege assumptions and increasing the chance of unexpected data collection or unauthorized task execution.

Description-Behavior Mismatch

Medium

• Confidence: 91% confidence
• Finding: This file implements keyword-based search task creation, which materially expands the skill beyond its declared purpose of routing and analyzing user-supplied Xiaohongshu links. That scope expansion can enable collection or enumeration of unrelated content based on arbitrary keywords, creating a data-access and capability mismatch that may violate user expectations and platform constraints.

Description-Behavior Mismatch

Medium

• Confidence: 90% confidence
• Finding: The code returns note and profile URLs derived from keyword search results rather than from links the user supplied. In the context of a link-detail skill, this turns the tool into a discovery/enumeration mechanism that can surface unrelated user content and profiles, increasing privacy and misuse risk beyond the declared workflow.

Description-Behavior Mismatch

Medium

• Confidence: 89% confidence
• Finding: The CLI persists the full fetched comment results to a local JSON log file without any clear user consent or disclosure. Because comment content may include personal data or sensitive context, this creates an unnecessary local data-retention surface that exceeds the stated link-analysis/routing behavior and can expose data to other local users, backup systems, or later unintended processing.

Context-Inappropriate Capability

Medium

• Confidence: 82% confidence
• Finding: The CLI persists the full fetched Xiaohongshu detail response to a local file after completing the request. Because the skill’s stated purpose is on-demand link analysis rather than data retention, this creates unnecessary storage of potentially sensitive or regulated content, increasing the risk of local data exposure, over-collection, and unintended reuse by other processes or users on the host.

Description-Behavior Mismatch

High

• Confidence: 92% confidence
• Finding: This CLI implements unrestricted keyword-based Xiaohongshu search, while the skill metadata says the skill should only route and analyze user-provided Xiaohongshu links. That scope expansion increases data access and capability beyond declared intent, which is dangerous in an agent setting because it enables broad content discovery and collection unrelated to a supplied URL.

Context-Inappropriate Capability

Medium

• Confidence: 87% confidence
• Finding: The code creates and retrieves a general search task from a keyword rather than deriving results from a user-supplied Xiaohongshu link. In the context of a link-analysis skill, this hidden or undeclared capability can be abused to perform unintended platform-wide searches, undermining least privilege and violating user/operator expectations about what the skill does.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The skill processes user-provided Xiaohongshu links, keywords, and related query context through a third-party API service, but the user-facing description does not clearly warn that this data will be transmitted off-platform. This creates a privacy and consent risk: users may share URLs containing tracking parameters or personal/business-sensitive research targets without realizing the data is sent to an external provider.

Missing User Warnings

Medium

• Confidence: 84% confidence
• Finding: The tool sends the user-supplied Xiaohongshu URL and limit parameter to an external service using an API token, but the user-facing behavior shown here does not explicitly warn that their input will be transmitted off-box. This can create privacy and trust issues, especially because URLs may embed identifiers or tokens and the skill description frames the feature as local link-based content retrieval rather than third-party processing.

Missing User Warnings

Medium

• Confidence: 90% confidence
• Finding: The program silently saves retrieved comment results to disk after displaying them, without informing the user that local persistence will occur. This increases privacy risk and data exposure because fetched comments may be stored beyond the user session, discoverable by other processes/users, or captured by system backups.

Static analysis

Detected: suspicious.exposed_secret_literal

File appears to expose a hardcoded API secret or token.

Critical

• Code: suspicious.exposed_secret_literal
• Location: SKILL.md:16