Security audit

Strategy Backtester

Security checks for vulnerabilities and agentic risk

Overview

This is a local CSV-based strategy backtesting helper with some overstated analysis claims, but no hidden access, network use, trading actions, or destructive behavior.

Install only if you want a local historical backtesting helper and can provide complete CSV data and assumptions. Treat results as historical simulations, not investment advice, and manually verify benchmark comparisons and bias warnings because those parts are not fully automated by the included script.

SkillSpector

Vulnerability Patterns

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

Findings (2)

Tp4

High

Category: MCP Tool Poisoning
Confidence: 92% confidence
Finding: The skill description materially overstates what the implementation appears to do and omits important side effects like writing output files. That mismatch can mislead downstream agents or users into trusting benchmark comparisons, bias warnings, and broader strategy validation that are not actually performed, creating unsafe decision-making based on incomplete or inaccurate analysis rather than direct code execution risk.

Vague Triggers

Medium

Confidence: 89% confidence
Finding: The skill enables implicit invocation without any trigger constraints, exclusions, or user-confirmation guardrails. That can cause the backtester to activate in contexts where the user did not clearly request financial backtesting, leading to unintended processing of strategy inputs and potentially over-trusting generated financial analysis.

Static analysis

No suspicious patterns detected.