guaikei-douyin-deep-dive-collector

• Outcome: Review
• Latest audit: Aug 25, 2026 · 4:32 AM UTC
• Version: 1.0.0

Overview

This Douyin data-collection skill is mostly purpose-aligned, but it should be reviewed because it can auto-trigger on vague research requests and persist large collected datasets locally by default.

Review before installing. Use it only when you intentionally want Douyin public-data collection, treat GUAIKEI_API_TOKEN as a secret, and avoid running it in shared workspaces unless you are comfortable with raw search, post, and comment results being saved under the skill's logs directory.

SkillSpector

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• 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 (11)

Description-Behavior Mismatch

Medium

• Confidence: 89% confidence
• Finding: The schema explicitly documents `video_url` as including both playback and download addresses, which conflicts with the stated skill boundary that it is not intended for downloading materials. Even though this file is only a schema, exposing and normalizing download URLs in the interface makes downstream misuse easier and weakens policy enforcement around content acquisition.

Description-Behavior Mismatch

Medium

• Confidence: 86% confidence
• Finding: The CLI writes fetched comment data to a local JSON file automatically after displaying results, creating persistent storage of potentially sensitive or regulated user-generated content. In this skill’s context, the manifest describes analysis/search of Douyin content, not local retention, so the extra data persistence expands data handling risk and can surprise users or downstream systems.

Description-Behavior Mismatch

Medium

• Confidence: 87% confidence
• Finding: The CLI persists the full fetched Douyin results to a local JSON file after completing a read-only collection task. This creates an undeclared data-retention behavior that can expose scraped content, account data, or analysis artifacts to other local users, backups, or downstream tooling, especially because users may reasonably expect inspection-only behavior from the skill description.

Vague Triggers

High

• Confidence: 95% confidence
• Finding: The description says the skill should be used even when the user has not mentioned Douyin, which broadens activation beyond clear user intent. Overly broad triggering can cause the agent to collect external platform data unexpectedly, leading to privacy, compliance, and user-trust issues through inappropriate tool invocation.

Vague Triggers

High

• Confidence: 96% confidence
• Finding: The generalized trigger extends to vague multi-step requests like 'help me research this track,' allowing the skill to self-select into broad workflows without sufficient constraints. In an agentic system, this can cause unanticipated data collection, scope creep, and execution of third-party API calls in contexts where the user did not specifically authorize Douyin scraping or analysis.

Missing User Warnings

Medium

• Confidence: 86% confidence
• Finding: The README explicitly says collected data is automatically exported to local JSON log files, but it does not prominently warn that these files may contain large volumes of scraped public data, comments, and identifiers that could create privacy, retention, or accidental-sharing risks. In this skill’s context—bulk collection and analysis of Douyin videos, authors, comments, and trends—automatic persistence increases the chance that sensitive or regulated data is stored longer than intended and exposed through backups, repo commits, or shared workspaces.

Missing User Warnings

Low

• Confidence: 81% confidence
• Finding: The setup instructions tell users to configure GUAIKEI_API_TOKEN during the main quick-start flow, but they do not clearly emphasize at that point that the token is a secret that must not be hardcoded, logged, pasted into shared shells, or committed to source control. Because this skill relies on a paid/privileged API token and encourages easy local execution, weak guidance can lead to credential leakage and unauthorized API use.

Missing User Warnings

Medium

• Confidence: 90% confidence
• Finding: The code saves comment results to disk without a clear user-facing warning or consent flow, which can expose collected comments to other local users, backup systems, or later reuse outside the original session. Because this is a comment-collection tool, silent persistence is more dangerous than in a purely transient viewer: users may assume they are only inspecting content, while the tool is actually retaining it.

Missing User Warnings

Low

• Confidence: 85% confidence
• Finding: The code writes fetched results to a local JSON file without an explicit warning or consent at the point of persistence. Even if the filename is sanitized, the silent write can surprise users and leak collected data into the local filesystem, shell environments, backups, or shared workspaces.

Missing User Warnings

Medium

• Confidence: 88% confidence
• Finding: The CLI writes the full search output, including user-provided keywords and returned results, to a local file without any visible consent, opt-in, or retention control. This creates a privacy exposure on shared systems or agent hosts because sensitive research topics, user interests, or collected platform data may persist on disk and be accessible to other users, backups, or later processes.

Missing User Warnings

Medium

• Confidence: 84% confidence
• Finding: The tool sends user-supplied search keywords to a remote API to perform the search, but the code does not provide an explicit privacy notice or confirmation at the point of transmission. In this skill context, remote transmission is functionally necessary, but undisclosed forwarding can still expose sensitive user interests, investigation topics, or business research terms to the service operator.

VirusTotal

VirusTotal findings are pending for this skill version.

Static analysis

No suspicious patterns detected.