Security audit

meta-fintech-engineer

Security checks for vulnerabilities and agentic risk

Overview

The skill is not malicious, but it asks agents to keep persistent self-reflection/learning records for fintech work without clear limits or user control.

Review before installing. Use this only if you are comfortable with a fintech skill keeping local learning notes, and avoid letting it record customer, account, transaction, credential, or proprietary business details unless the storage policy is clarified or disabled.

SkillSpector

Vulnerability Patterns

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
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (2)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 86% confidence
Finding: The skill text explicitly describes writing to a learner/self-reflection loop and persistent evolution state, which implies file-write capability without any declared permission boundary or user-consent model. Undeclared persistence is risky because it can store sensitive task content, outcomes, or operator notes locally in ways users do not expect or control.

Tp4

High

Category: MCP Tool Poisoning
Confidence: 91% confidence
Finding: The declared purpose presents the skill as a fintech meta-skill, but the behavior includes persistent local logging and recording free-form notes/failure statistics that are not transparently disclosed as operational side effects. In a fintech context, this mismatch is especially dangerous because reflected failures, outcomes, or notes may contain sensitive financial, customer, or internal decision data that gets retained unexpectedly.

Static analysis

No suspicious patterns detected.