guaikei-douyin-pull-comments-stream

• Outcome: Review
• Latest audit: Aug 21, 2026 · 1:23 AM UTC
• Version: 1.0.0

Overview

This skill mostly does what it says, but it can trigger on vague requests and automatically stores collected Douyin data locally without strong user controls.

Install only if you are comfortable sending Douyin search terms, video/account URLs, and the configured API token to GuaiKei, and with results being saved under the skill's logs/ directory by default. Use it for explicit Douyin research requests, and periodically delete logs if they contain sensitive topics or comment/user data.

SkillSpector

Vulnerability Patterns

• 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
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration

Findings (9)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 92% confidence
• Finding: The skill requires the environment variable `GUAIKEI_API_TOKEN`, which is a sensitive credential capability, but no explicit permission declaration is surfaced in the skill metadata. This creates a transparency and consent gap: users or hosting systems may invoke the skill without realizing it consumes secrets from the runtime environment.

Description-Behavior Mismatch

Medium

• Confidence: 89% confidence
• Finding: The schema explicitly describes `video_url` as both playback and download addresses, which contradicts the skill description stating it does not handle downloading videos. This mismatch can cause downstream agents or integrators to expose or rely on download functionality that the published skill contract claims to forbid, increasing the risk of policy bypass, compliance issues, or unreviewed data handling.

Description-Behavior Mismatch

Medium

• Confidence: 90% confidence
• Finding: The CLI persists fetched comment results to a local JSON file after printing them, even though the skill description frames the behavior as transient search/fetch/query functionality. Comment data can contain personal data or sensitive content, and silent local retention increases the risk of unintended disclosure, over-collection, and misuse on shared or monitored systems.

Vague Triggers

High

• Confidence: 95% confidence
• Finding: The top-level description says the skill should also apply when users say vague phrases like '查一下' or '帮我看看' even without mentioning Douyin. That broad trigger can cause unintended activation on generic requests, sending unrelated user queries to an external data collection workflow and potentially disclosing user intent or causing surprise actions.

Vague Triggers

Medium

• Confidence: 91% confidence
• Finding: The generalized activation rule states that any short-video content research, competitor analysis, or topic planning may trigger the skill even if the user never mentioned Douyin. In context, this increases the chance of cross-context invocation and accidental collection of third-party platform data assumptions, which is risky because the skill performs external lookups and logging.

Missing User Warnings

Medium

• Confidence: 96% confidence
• Finding: The skill states that logs are automatically written to `logs/` with filenames derived from timestamps, keywords, and type, but the description does not clearly warn users that their query inputs, URLs, and retrieved data may be stored locally. This can expose sensitive research topics, watched accounts, or analyzed videos to other local users, backups, or later unintended reuse.

Missing User Warnings

Medium

• Confidence: 90% confidence
• Finding: The README states that scraped data is automatically exported to local JSON log files, including comments and account/video-related data, but does not prominently warn users about privacy, retention, or access-control implications of storing potentially sensitive scraped content on disk. In this skill context, automatic persistence increases risk because the tool is explicitly used for monitoring public content and comments at scale, which can accumulate personal data and make accidental disclosure more likely.

Missing User Warnings

Medium

• Confidence: 92% confidence
• Finding: The code writes full comment results to a local JSON file without any explicit warning or consent mechanism in this file. Silent persistence of third-party comments can surprise users and leak collected data to other local users, backups, or malware, especially because filenames are predictable and derived from the request context.

Missing User Warnings

Medium

• Confidence: 84% confidence
• Finding: The CLI persists full search results to a local JSON file without obtaining explicit user consent or providing a clear warning. Because this skill handles search terms and scraped content that may be sensitive or operationally confidential, local persistence can create unintended data retention and disclosure risk on shared systems or multi-user environments.

VirusTotal

VirusTotal findings are pending for this skill version.

Static analysis

No suspicious patterns detected.