Security audit

股票分析雷达专业版

Security checks for vulnerabilities and agentic risk

Overview

This finance skill is mostly aligned with market analysis, but it overreaches into unrelated security wording and gives broad read/write/execute plus external-notification authority without enough scoping.

Review before installing. Use it only for finance workflows, avoid relying on the unrelated security claims, and require explicit confirmation before exports, broker/API sync, webhook/email notifications, callback URLs, or long-running monitoring. Do not provide portfolio files or API keys unless you trust the skill publisher and have checked where generated files and notifications will go.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (6)

Intent-Code Divergence

Medium

Confidence: 95% confidence
Finding: The manifest describes a finance-analysis skill but also claims it should be used for security detection, compliance audit, quality inspection, and cryptographic protection. This capability mismatch can cause the agent to invoke a finance-oriented skill in security-sensitive contexts where its behavior, outputs, and safeguards are not appropriate, increasing the risk of incorrect trust, unsafe execution, or misleading results.

Intent-Code Divergence

Medium

Confidence: 94% confidence
Finding: The capability coverage statement explicitly lists security detection, compliance auditing, and unauthorized-target language even though the rest of the document describes stock, crypto, portfolio, alerting, and rumor-scanning features. Overstating capability coverage can trigger the skill in unrelated high-trust workflows and mislead users or orchestrators into treating financial tooling as a security tool.

Vague Triggers

High

Confidence: 96% confidence
Finding: The description contains broad trigger text such as security detection, compliance audit, quality checking, and cryptographic protection despite the skill being categorized as finance. Ambiguous activation scope is dangerous because agent systems may route unrelated or sensitive requests into a tool with exec/write/network capabilities, leading to inappropriate execution paths or overcollection of data.

Vague Triggers

High

Confidence: 95% confidence
Finding: The keyword-style coverage list is excessively broad and includes fragmented phrases that could match many unrelated prompts. In an agent environment, such loose matching materially raises the chance of unintended invocation of a skill that can read, write, and execute commands, which amplifies the security risk beyond a simple documentation issue.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The skill documents export and file-writing behavior such as CSV/Excel/JSON output without clearly warning that local files may be created or overwritten. In an agent setting, insufficient disclosure around write operations can lead to unintended filesystem modification, data sprawl, or overwriting user files when parameters are inferred automatically.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The skill supports webhook, email, and external API integrations but does not clearly warn that portfolio holdings, watchlists, or monitoring data may leave the local system. Because this skill handles financial context and may transmit user-specific data, lack of a privacy and egress warning can result in unintentional disclosure to third parties.

Static analysis

No suspicious patterns detected.