guaikei-douyin-competitor-intel-source

• Outcome: Review
• Latest audit: Aug 24, 2026 · 2:33 AM UTC
• Version: 1.0.0

Overview

The skill mostly does what it says, but it can activate on overly broad search requests, sends queries to a third-party API, and automatically stores retrieved Douyin data locally.

Install only if you want an API-backed Douyin research tool. Be aware that search terms and Douyin URLs are sent to guaikei.com with your token, and successful search/comment/account runs save full JSON logs locally. Use it only when the user has clearly asked for Douyin data, and clean up logs if they contain sensitive research topics or personal data.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands

Findings (10)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 82% confidence
• Finding: The skill explicitly relies on an environment variable (`GUAIKEI_API_TOKEN`) but does not declare permissions in a structured way, creating a mismatch between documented behavior and exposed capabilities. This can weaken reviewability and least-privilege enforcement, making it easier for a skill to access secrets without clear operator awareness.

Description-Behavior Mismatch

Medium

• Confidence: 92% confidence
• Finding: The CLI writes the full fetched comment results to a local JSON file after retrieval, which goes beyond a purely read-only/query operation and creates a persistent local copy of potentially sensitive third-party data. This increases privacy and data-retention risk because comments may contain personal information and the file may be accessible to other local users, later processes, or backups without the user's awareness.

Vague Triggers

High

• Confidence: 94% confidence
• Finding: The description instructs the agent to activate even when the user did not mention Douyin, using very broad phrases like '帮我搜下' or '网上谁在讲这个'. This can cause overbroad routing of unrelated research requests into a networked data-collection skill, leading to unintended external queries, privacy issues, and surprising behavior outside the user’s intended platform.

Vague Triggers

Medium

• Confidence: 91% confidence
• Finding: The generalized activation guidance says multi-step research tasks should trigger this skill even without explicit mention of Douyin, which expands the skill’s reach beyond clearly bounded intent. In context, this is more dangerous because the skill performs external data collection and uses an API token, so mistaken activation can leak sensitive research terms or cause unauthorized third-party querying.

Vague Triggers

Medium

• Confidence: 93% confidence
• Finding: The natural-language trigger examples are broad enough to match common phrases like '帮我搜下' or '网上谁在讲这个' even when a user did not clearly request Douyin scraping. That can cause unintended activation of a data-collection skill, leading to over-collection of platform data, comment retrieval, and local log creation without sufficiently explicit user intent.

Vague Triggers

Medium

• Confidence: 91% confidence
• Finding: The intent-recognition rules rely on generic terms such as '热搜', '搜索', and '找 + 关键词', which overlap with ordinary assistant requests and create ambiguous routing conditions. In this skill, ambiguous routing is more dangerous because execution invokes external scraping-style collection and may persist retrieved data to disk, so a misfire is not just a harmless misunderstanding.

Missing User Warnings

Medium

• Confidence: 89% confidence
• Finding: The README promotes comment and account-data retrieval but does not prominently warn users that these workflows can collect privacy-sensitive public data at scale and store it in local JSON logs. Even if the source data is public, bulk aggregation plus local persistence increases privacy, compliance, and accidental disclosure risk for operators and affected individuals.

Missing User Warnings

Low

• Confidence: 89% confidence
• Finding: The tool silently writes retrieved comment data to disk without explicit warning or user choice. Even though the filename is sanitized, the behavior can expose scraped content to unintended local access, backups, or downstream tooling, making it a real privacy and least-surprise issue.

Missing User Warnings

Low

• Confidence: 90% confidence
• Finding: The CLI persists full search output to a local file named with the user's keyword, which can expose potentially sensitive search terms and retrieved content to other local users, backups, or later forensic review without an obvious user-facing warning or opt-in. In this skill's context, competitor-intelligence and comment-search queries may contain commercially sensitive topics, making silent persistence a real privacy risk even if it is not a code-execution issue.

Missing User Warnings

Low

• Confidence: 84% confidence
• Finding: The program sends user-supplied keywords and filters to remote API calls without any explicit privacy notice or consent flow. Because this skill is specifically used for searching Douyin content and competitor intelligence, user queries may reveal business strategy, investigations, or sensitive interests to the service operator or intermediaries.

VirusTotal

VirusTotal findings are pending for this skill version.

Static analysis

No suspicious patterns detected.