Back to skill

Security audit

CFO Global Super Expert System

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed CFO-themed router and index for other finance/business skills, with no evidence of hidden persistence, exfiltration, or destructive behavior.

Before installing, treat this as a routing/index skill rather than an authoritative CFO, legal, audit, or investment advisor. Review any sub-skills it loads, especially stock-master-hunter and other external dependencies, before letting them access real-time data or produce advice for consequential decisions.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • 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
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (6)

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The skill is presented as a full CFO expert system, but the described behavior is mainly routing to other skills and local utilities rather than providing the claimed integrated expertise. This mismatch can mislead users and downstream agents into over-trusting outputs, invoking unnecessary local resources, or assuming validated professional capability where none is implemented.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
84% confidence
Finding

The routing rules use broad generic terms like 'market' and 'industry' that can match many ordinary conversations, causing unintended activation of this skill. Overbroad triggering can route unrelated prompts into finance/investment workflows, increasing the chance of irrelevant tool use, misleading advice, or unnecessary access to linked sub-skills.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The keyword map includes highly generic business terms such as 'audit', 'tax', 'strategy', 'compliance', 'leadership', and 'industry' without clear scope boundaries. In a meta-router that loads other skills and may invoke local resources, this can cause excessive or incorrect delegation, broadening the attack surface and creating trust confusion about which specialized component is acting.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

The manifest frames the skill as a CFO-level expert system for financial, investment, governance, audit, strategy, and presentation analysis. The documented capability to invoke python3 {baseDir}/scripts/cfo_router.py ... is an operational tooling feature for local script execution rather than a direct finance-domain function, and the manifest does not explicitly declare command execution as part of scope.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
79% confidence
Finding

This Python skill emits user-facing text such as headings, usage strings, and status labels only in English, even though the registry content includes both English and Chinese terms. That creates a locale policy concern because the skill selects a language unilaterally and does not provide any opt-in or language-selection mechanism.

Content

No source excerpt is available for this finding.