Security audit

A股每日新闻

Security checks across malware telemetry and agentic risk

Overview

This A-share sentiment skill is mostly purpose-aligned, but it needs Review because it can send finance queries to external services, use local API credentials, create local reports, and auto-open generated HTML without clear user confirmation.

Install only if you are comfortable sending stock research terms, watchlists, and related market queries to RedFox and web search providers. Use a revocable API key, avoid nonpublic portfolio or trading-plan details, and review generated HTML files before opening them automatically.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands

Findings (6)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 89% confidence
Finding: The skill declares powerful tool access and operational behavior that reads environment variables, performs network-backed searches, and writes local files, yet there is no explicit permissions model or user-facing gating for those capabilities. In this context, the skill also directs automatic report generation and local file output, so the lack of declared permissions reduces transparency and increases the chance of users authorizing broader access than they understand.

Context-Inappropriate Capability

Medium

Confidence: 95% confidence
Finding: The output template hardcodes promotional links to other subscription skills that are not necessary to fulfill the user’s requested A-share news analysis. This creates a cross-skill steering path that can manipulate users into unrelated follow-on actions, increasing the risk of unsolicited tracking, spammy upsells, or privilege expansion beyond the original task.

Vague Triggers

Medium

Confidence: 86% confidence
Finding: The example trigger phrases are broad enough to match common user requests such as '看看最新数据' or generic market/news queries, which can cause the skill to activate unintentionally. Because this skill sends queries to external services and generates reports, accidental invocation can disclose user query content to third parties and perform actions the user did not clearly intend.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The README mentions an API key but does not clearly warn that user queries and possibly derived search terms will be transmitted to external third-party platforms/APIs. In a finance-related skill, queries may contain sensitive research interests, trading intent, or proprietary watchlists, so missing disclosure undermines informed consent and increases privacy/compliance risk.

Vague Triggers

Medium

Confidence: 84% confidence
Finding: The trigger list contains broad terms such as A股, 选股, 股票讨论, and 大盘分析, which are common in ordinary finance conversations and can cause the skill to invoke when the user did not intend it. Because the skill performs web searches, calls a script, and generates files, accidental invocation can lead to unnecessary network activity, local file creation, and unintended use of configured API credentials.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill instructs the agent to automatically generate an HTML report and then open it locally without asking the user. Auto-opening a generated HTML file is risky because HTML can contain active content or dangerous links, and this behavior crosses from passive file creation into execution-like system interaction without explicit consent; in a skill that aggregates external content, the risk is heightened because untrusted data may be rendered in the report.

VirusTotal

62/62 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.