Security audit

Stock Evaluator

Security checks for vulnerabilities and agentic risk

Overview

This stock-analysis skill is mostly coherent, but it gives concrete buy/sell and position-sizing guidance without mandatory financial-advice limits and includes realistic sample dashboard data that could be mistaken for real market data.

Review this carefully before installing if you may rely on it for real-money decisions. It can produce highly specific investment actions and portfolio allocations, so users should independently verify every number, treat outputs as informational analysis rather than financial advice, and ensure any dashboard values are replaced with sourced current data rather than template examples.

A.I.G

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

SkillSpector

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (10)

Missing User Warnings

High

Confidence: 98% confidence
Finding: This workflow instructs the agent to provide concrete investment actions, entry/exit prices, stop-losses, and position sizing without any mandatory risk disclosure, suitability check, or limitation framing. In a stock-picking skill, that context makes the issue more dangerous because users are likely to rely on the output for real-money decisions, creating a material risk of financial harm from inaccurate, unsuitable, or overconfident recommendations.

Intent-Code Divergence

High

Confidence: 97% confidence
Finding: The skill explicitly mandates a zero-fabrication policy, yet the embedded React template contains realistic-looking sample metrics, news headlines, price histories, and recommendation fields that can be copied through into output as if they were current facts. In a finance skill, this creates a strong path for fabricated or stale market data to reach end users, leading to materially misleading investment analysis and potentially harmful financial decisions.

Ae4

Medium

Category: analysis-evasion
Confidence: 80% confidence
Finding: Suspicious Unicode normalization or mixed-script content

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: This guide gives action-oriented investment screening guidance such as avoiding companies and selecting quality stocks without clearly warning that the material is informational only and should not be relied on as personalized financial advice. In the context of a stock-evaluation skill that produces recommendations, users may over-trust these outputs and make harmful financial decisions without independent verification.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The text gives strong prescriptive conclusions like 'Strong candidate' and 'Avoid' without an explicit nearby warning about financial loss risk, uncertainty, or need for independent verification. Because this skill is explicitly designed to provide entry prices, position sizing, and conviction ratings, such definitive language increases the chance that users treat the content as authoritative investment advice.

Natural-Language Policy Violations

Medium

Confidence: 92% confidence
Finding: This markdown file contains a natural-language locale constraint: 'Professional styling with € currency default.' The document later reinforces that all monetary values are in Euro by default, but does not provide opt-in, alternatives, or a region-specific justification, which can violate language/locale policy expectations.

Natural-Language Policy Violations

Medium

Confidence: 97% confidence
Finding: The section explicitly states that all monetary values are in Euro by default and gives all examples in that currency. Because this applies globally and does not offer a user choice or explain a region-specific compliance need, it is a natural-language locale policy concern.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill is designed to produce explicit buy/hold/sell calls, entry prices, stop losses, and portfolio position sizing, but it does not require a clear user-facing notice that the output is not financial advice and may be incomplete or wrong. This increases the chance that users over-rely on the tool for consequential financial decisions without understanding its limitations.

Intent-Code Divergence

Medium

Confidence: 99% confidence
Finding: The instructions say missing insider data must be displayed as "N/A" and must never be substituted with zero, but the template defaults insider buys and sells to 0. In this context, zero has semantic meaning and can falsely signal absence of insider activity, distorting sentiment assessment and investment conclusions.

Natural-Language Policy Violations

Low

Confidence: 98% confidence
Finding: The markdown instructs that all monetary values should be displayed in Euro by default, which imposes a locale/currency preference on all users. Under the language/locale policy, forced locale conventions without user choice should be flagged unless clearly justified as region-specific, which is not stated here.

Static analysis

No suspicious patterns detected.