Security audit

快手洞察与竞品分析助手(KuaiShou Data Miner)

Security checks for vulnerabilities and agentic risk

Overview

This skill coherently retrieves user-requested public Kuaishou data through a disclosed third-party API and saves results locally, with no evidence of hidden access, destructive behavior, or account takeover.

Install only if you are comfortable sending Kuaishou search terms, profile/video URLs, and the GUAIKEI_API_TOKEN to the guaikei.com API. Review or delete the generated logs directory when results include competitor research, user comments, or other data you do not want retained on disk.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (6)

Description-Behavior Mismatch

Medium

Confidence: 87% confidence
Finding: The script persists fetched comment data to a local file after completing the requested retrieval, which creates a secondary data-retention behavior beyond simply returning results to stdout. Even if the data is public, local persistence can expose collected comments, URLs, and metadata to other local users, backup systems, or later unintended reuse without the user's awareness.

Description-Behavior Mismatch

Low

Confidence: 92% confidence
Finding: The CLI writes the full search output to a local JSON file whose filename includes the user keyword, causing both query terms and returned data to persist on disk beyond the immediate command output. In a data-mining skill that may process competitor research, creator data, or comment data, this creates an unnecessary local data-retention trail that can leak through shared machines, backups, logs, or later file collection.

Missing User Warnings

Medium

Confidence: 84% confidence
Finding: The README states that all task results are automatically saved to the `logs/` directory, but it does not prominently warn users that search terms, creator URLs, comment data, and analysis outputs may persist on disk. In a data-mining skill that handles public-but-potentially sensitive profiling and comment datasets, silent persistence increases the risk of unintended retention, local disclosure, and mishandling of scraped data.

Missing User Warnings

Medium

Confidence: 89% confidence
Finding: The code writes comment results to local storage without any explicit warning, confirmation, or user-controlled save path. Silent persistence can surprise users and create privacy, compliance, and data-governance issues, especially when comments may contain personal information or sensitive discussion content.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The CLI writes the full fetched result set to a local JSON file automatically after successful execution, without any opt-in, path control, or warning. Because this skill handles creator data and comments at scale, the output may contain sensitive or regulated public-profile data that can persist on disk, be exposed to other local users/processes, or be unintentionally committed or exfiltrated later.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: Search results are silently persisted without explicit user notice, while the stored content includes the request metadata and full results payload. Because this skill is specifically designed for large-scale public-data retrieval and analysis, the saved files can contain sensitive research terms, creator targets, and comment datasets, increasing privacy, compliance, and operational exposure if the workstation or repository is later accessed by others.

Static analysis

No suspicious patterns detected.