guaikei-douyin-live-data-puller

• Outcome: Review
• Latest audit: Aug 25, 2026 · 4:18 AM UTC
• Version: 1.0.0

Overview

This Douyin data-pulling skill largely does what it claims, but it needs review because it can activate on overly broad requests and saves collected social-media data locally by default.

Install only if you are comfortable sending search terms, Douyin URLs, and the GUAIKEI_API_TOKEN to guaikei.com, and with result files being saved under the skill logs directory by default. Use it only for explicit Douyin public-data tasks, avoid sensitive queries on shared machines, and delete exported logs when no longer needed.

SkillSpector

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• 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

Findings (13)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 88% confidence
• Finding: The skill explicitly depends on the environment variable `GUAIKEI_API_TOKEN`, but the metadata does not declare permissions for environment access. This creates a transparency and least-privilege problem: callers and reviewers cannot accurately assess what sensitive inputs the skill needs, increasing the risk of unintended secret exposure or unsafe execution assumptions.

Description-Behavior Mismatch

Medium

• Confidence: 94% confidence
• Finding: The schema explicitly documents `video_url` as both playback and download addresses, which conflicts with the skill’s stated scope excluding download-related use cases. This can enable downstream agents or integrators to treat download URLs as an approved output and build workflows that facilitate prohibited media downloading.

Description-Behavior Mismatch

Medium

• Confidence: 89% confidence
• Finding: The CLI persists fetched comment data to a local JSON file after presenting itself as a retrieval/browsing tool, creating undisclosed local data retention. Because comments may contain personal data or sensitive content, this expands the data handling surface and can expose user activity or scraped content to other local users, backup systems, or downstream tooling.

Context-Inappropriate Capability

Medium

• Confidence: 96% confidence
• Finding: The CLI persists the full fetched Douyin results to a local JSON file after completing the request, even though the user-facing purpose is only to search/view content. Silent local retention increases privacy and data-handling risk because scraped post data may include identifiers, metadata, or content the user did not expect to be stored, and those files can later be accessed by other local users, processes, backups, or tooling.

Description-Behavior Mismatch

Medium

• Confidence: 92% confidence
• Finding: The CLI writes the full search output, including user-supplied keyword and returned results, to a local file without any visible consent, warning, or opt-in. In a search/browsing skill, this creates unnecessary data persistence and can expose sensitive queries or collected content to other local users, backup systems, or later compromise of the host.

Vague Triggers

High

• Confidence: 95% confidence
• Finding: The description instructs activation on very broad phrases like '帮我搜下', '网上谁在讲这个', and '这个账号更新了啥' even when the user never mentions Douyin. That can cause the agent to invoke this third-party data-pulling skill for generic research requests, leading to unintended data access, wrong-platform actions, and over-collection beyond user intent.

Vague Triggers

High

• Confidence: 94% confidence
• Finding: The generalized activation guidance says the skill should trigger for unspecified 'short video research', 'competitive analysis', and buried multi-step workflows even without naming Douyin. In a larger agent workflow, this can silently expand the skill's scope and cause hidden tool invocation, making user intent mismatches and unnecessary external API use more likely.

Missing User Warnings

Medium

• Confidence: 89% confidence
• Finding: The README states that scraped results are automatically exported to local JSON logs, including potentially sensitive comment, account, and activity data, but it does not prominently warn that this persistence happens by default. In a data-collection skill, silent on-disk retention increases the risk of unintended disclosure through shared machines, backups, source control, or later reuse of the files.

Missing User Warnings

Low

• Confidence: 76% confidence
• Finding: The README instructs users to place the API token in an environment variable before use, but the credential-handling warning appears later and is not prominent. While environment variables are common, insufficient guidance can still lead to token exposure via shell history, screenshots, shared sessions, CI logs, or misconfigured environments.

Missing User Warnings

Medium

• Confidence: 88% confidence
• Finding: The document operationalizes collection of creator posts and comment data at scale, including up to 10,000 records, but provides no privacy, consent, retention, or terms-of-service guidance. In a skill explicitly designed to pull Douyin content, this omission increases the likelihood of misuse for profiling, harassment, surveillance, or non-compliant scraping of user-generated data.

Missing User Warnings

Medium

• Confidence: 92% confidence
• Finding: The program writes full comment results to a local JSON file without any user-facing warning, consent, or obvious need from the skill description. Silent persistence of scraped social-media comments can surprise users and lead to privacy, compliance, or workstation data-leak issues, especially on shared systems or in automated agent environments.

Missing User Warnings

Medium

• Confidence: 98% confidence
• Finding: This code writes fetched results to a local JSON file without any user warning, confirmation, or explicit save option. Undisclosed persistence is dangerous because operators may believe the tool only prints transient output, while in reality it creates an additional data copy that can expose sensitive browsing targets, collected content, and account-related metadata through the filesystem.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: Search results are persisted locally with no clear user warning that queries and retrieved content will be stored on disk. Because this skill is for browsing/searching potentially sensitive topics or accounts, silent retention increases privacy risk and may violate user expectations or organizational data-handling requirements.

VirusTotal

VirusTotal findings are pending for this skill version.

Static analysis

No suspicious patterns detected.