Security audit

guaikei-xhs-note-and-comment

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed Xiaohongshu public-data fetcher with local result logging that users should be aware of before use.

Install only if you are comfortable sending Xiaohongshu keywords or URLs and your GUAIKEI_API_TOKEN to guaikei.com. Treat fetched comments and creator data as potentially personal even if public, and periodically clear the skill's logs directory if you do not want retained local copies.

SkillSpector

By NVIDIA

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (4)

Context-Inappropriate Capability

Medium

Confidence: 88% confidence
Finding: The CLI persists fetched comment results to a local file after already printing them to stdout, creating an additional data sink that users may not expect. Even if the source data is public, comments can still contain personal information or sensitive context, and silent local retention increases the chance of accidental disclosure, over-collection, or insecure storage on shared systems.

Context-Inappropriate Capability

Medium

Confidence: 91% confidence
Finding: The CLI persists full search output to a local file, which creates an unintended data-retention channel outside the user-visible response path. Even if the source data is nominally public, stored query terms and result sets can expose user interests, operational history, or sensitive research topics, and the manifest does not make this persistence behavior clear.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The code writes comment results to a local JSON file without any user-facing notice or confirmation, which can surprise users and leave scraped data behind on disk. In CLI and agent contexts, silent persistence is risky because operators may assume the tool only returns data transiently, while the saved file can later be accessed by other users, backups, or unrelated processes.

Missing User Warnings

Low

Confidence: 89% confidence
Finding: Writing search results to a local file without an explicit user-facing warning reduces transparency and can surprise operators who expect a fetch-only public-data skill. The main risk is undisclosed persistence of user queries and returned content, which can accumulate sensitive context over time even when the upstream data is public.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

Detected: suspicious.exposed_secret_literal

File appears to expose a hardcoded API secret or token.

Critical

Code: suspicious.exposed_secret_literal
Location: SKILL.md:16