ClawHub
Back to skill

Security audit

财务分析专业套件

Security checks across malware telemetry and agentic risk

Overview

The skill is a finance-analysis helper, but it asks for broad command/file authority while giving overly broad activation and inconsistent API-key storage guidance.

Review this before installing if your workspace contains sensitive files or credentials. Use environment variables or a secret manager for API keys, avoid putting live keys in config.yaml, and invoke the skill only for explicit finance/accounting analysis tasks where read/write/exec access is necessary.

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 (4)

Intent-Code Divergence

Medium
Confidence
95% confidence
Finding
The skill states both that API keys should be provided via environment variables and that keys are stored in a local config.yaml file, which creates ambiguous credential-handling guidance. In a skill with read/write/exec capabilities and external data-source integration, this ambiguity can lead users to persist secrets in plaintext configuration files that may be exposed, committed, or read by tooling.

Vague Triggers

Medium
Confidence
89% confidence
Finding
The activation language is broad enough to match generic requests about efficiency, automation, workflow optimization, and batch processing, which can cause the skill to trigger outside intended finance-specific contexts. Because the skill advertises exec, read, and write capabilities, overbroad activation increases the chance that it is invoked in unrelated tasks where file access or command execution is unnecessary and risky.

Vague Triggers

Medium
Confidence
90% confidence
Finding
The capability coverage keywords include vague trigger phrases such as efficiency improvement, automation, workflow optimization, and enterprise team usage without tight scope controls. In context, this broad matching is more dangerous because the skill is positioned as an agent-executable workflow and may gain access to commands, files, or external integrations based on weak intent matching.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The documentation instructs users to copy a template config and fill in data-source credentials in config.yaml without a strong warning against storing secrets in plaintext. This is especially risky in a skill that supports file operations, automation, and reporting because secrets in local config files are more likely to be read by tools, leaked in logs, or committed to version control.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.