Security audit

基金监控

Security checks for vulnerabilities and agentic risk

Overview

This fund-monitoring skill is mostly coherent, but it ships with a preconfigured Feishu webhook that could send fund reports to an unknown channel.

Review before installing. Remove the bundled Feishu webhook from data/config.json and data/push_channels.json, then configure only your own notification destinations. Avoid entering cost basis, share counts, or other sensitive holdings data until you confirm where reports will be sent and how exported files will be stored.

SkillSpector

Vulnerability Patterns

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

Findings (3)

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: The skill explicitly documents exporting task details and holdings-related monitoring data to local files, including a default Windows path, but does not warn users that these files may contain sensitive financial information and could be exposed through shared machines, backups, malware, or weak filesystem permissions. Because the feature is framed as convenient archiving, users may enable it without understanding the confidentiality risk.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The skill supports sending monitoring reports and holdings/profit-loss information to third-party services such as WeCom, DingTalk, Feishu, and SMTP email, but it does not clearly warn that potentially sensitive financial data will leave the local environment and be processed by external providers. This creates a privacy and data-handling risk, especially if users paste personal webhooks or email credentials without understanding the exposure.

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: The code can send monitoring reports, including holdings-derived profit/loss information, to externally configured webhook and email endpoints with no validation, redaction, or explicit confirmation at send time. In this skill's context, that creates a real confidentiality risk because personal financial positions and watchlists may be disclosed to third-party services or misconfigured destinations.

Static analysis

No suspicious patterns detected.