Guaikei Douyin Search And Analyze

• Outcome: Review
• Latest audit: Aug 19, 2026 · 1:47 AM UTC
• Version: 1.0.0

Overview

The skill is broadly coherent, but it needs review because it can run on ambiguous short-video requests, sends queries and tokens to a third-party service, and automatically saves full scraped results locally.

Review before installing if you work with sensitive research topics, shared machines, or regulated personal data. Treat all results as data retained on disk under the skill's logs directory, and assume keywords, target URLs, limits, and the API token are sent to Guaikei's service.

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 Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands

Findings (11)

Description-Behavior Mismatch

Medium

• Confidence: 86% confidence
• Finding: The CLI persists full fetched comment results to a local JSON log file after completing the request. Because comments may contain personal data, usernames, or other sensitive content, this creates unnecessary at-rest data retention beyond the skill’s stated search/read behavior and increases exposure if the host filesystem is shared, synced, or later inspected.

Description-Behavior Mismatch

Medium

• Confidence: 95% confidence
• Finding: The CLI persists the fetched Douyin post results to a local JSON file after completing a read/analyze operation, which expands the data lifecycle beyond what the skill description communicates. Even if the source data is public, local persistence can unexpectedly retain scraped content, metadata, and query targets on disk where other local users, tools, or later processes may access it.

Description-Behavior Mismatch

Medium

• Confidence: 94% confidence
• Finding: The CLI writes the full search output and request metadata to a local file, including user-supplied keywords and returned content, without any explicit consent, visibility, retention control, or opt-out. In a search/analysis skill, this creates an unnecessary data persistence channel that can expose user activity and collected platform data to other local users, backup systems, or later compromise of the host.

Context-Inappropriate Capability

Medium

• Confidence: 86% confidence
• Finding: This utility accepts a caller-controlled filename and content, then writes a file to local disk. Although it applies basic filename sanitization and constrains writes to a logs directory, it still gives the skill persistent local file-write capability that is not clearly required for a Douyin search/analysis skill, increasing the risk of sensitive data retention, unintended artifact creation, or abuse by other code paths.

Vague Triggers

Medium

• Confidence: 91% confidence
• Finding: The activation description is intentionally broad, stating the skill should be used for general 'short-video platform' style questions even when the user does not explicitly mention Douyin. This can cause the agent to invoke an external data-collection skill in cases where the user intended another platform or only wanted general advice, leading to unintended external queries and possible over-collection of third-party data.

Vague Triggers

Medium

• Confidence: 94% confidence
• Finding: The usage guidance explicitly tells the agent to use the skill even when the user did not mention Douyin, which weakens consent and intent boundaries around external lookups. In context, this skill performs searches, account scraping of public works, comments retrieval, and hotlist checks, so an ambiguous trigger can cause unnecessary third-party API use and collection of data beyond the user's intended scope.

Missing User Warnings

Medium

• Confidence: 87% confidence
• Finding: The README promotes automatic JSON log export of scraped Douyin data for analysis and reporting, but it does not prominently warn that those logs may contain personal data, comments, account identifiers, or other regulated content. In a scraping/analytics skill, silent local retention and secondary use of collected data increases the risk of privacy violations, over-collection, and accidental disclosure if logs are shared or stored insecurely.

Missing User Warnings

Medium

• Confidence: 84% confidence
• Finding: The README instructs users to set GUAIKEI_API_TOKEN and use a third-party service, but the warning is limited to functionality and token setup rather than clearly disclosing that requests and user-supplied targets/queries will be sent to an external provider. This can lead operators to expose credentials or sensitive search targets without understanding the trust boundary or the need for secret-handling controls.

Missing User Warnings

Medium

• Confidence: 88% confidence
• Finding: The CLI sends the user-supplied Douyin URL together with an access token to remote API functions without clearly disclosing that this input will be transmitted off-host. In a skill marketed primarily as search/analysis, lack of explicit notice weakens informed consent and can expose queried targets, operator behavior, and credentials to a backend service unexpectedly.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The tool silently writes fetched results to a local JSON file without advertising that behavior in its help or interface, creating unexpected local data retention. This can leak browsing targets or scraped datasets to shared systems, backups, or later processes, especially because the filename is derived from the queried URL and timestamp.

Missing User Warnings

Medium

• Confidence: 96% confidence
• Finding: Search results are silently written to disk after successful execution, but the tool does not clearly warn the user that their query and collected results will be saved locally. This can surprise users and leak potentially sensitive research interests or collected data through local filesystem access, shared workstations, or routine log collection.

Static analysis

No suspicious patterns detected.