Wall Street Quant Team

• Outcome: Review
• Latest audit: May 28, 2026 · 1:10 PM UTC
• Version: 1.0.1

Overview

This skill is mostly an investment-analysis prompt pack, but it also describes automatic monitoring, trade execution, broad logging, and cross-skill loading in ways users should review carefully.

Install only if you will use it as decision-support content, not as a trading bot. Do not connect it to brokerage accounts, allow automatic stop-losses, enable scheduled monitoring, load extra advisor skills, or retain raw portfolio and strategy logs unless you explicitly approve and can control each behavior.

SkillSpector

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection

Findings (14)

Intent-Code Divergence

Medium

• Confidence: 92% confidence
• Finding: The documentation contradicts itself on whether the skill only provides investment suggestions or actually performs trading-related execution. In a financial workflow, ambiguity about execution authority is dangerous because downstream agents or integrators may interpret this as permission to place or close positions without proper user consent or controls.

Vague Triggers

Medium

• Confidence: 92% confidence
• Finding: The skill description and activation framing are broad enough that ordinary requests about stocks, market analysis, or team configuration could trigger the skill unexpectedly. Over-broad activation can route users into a domain-specific workflow without clear consent, increasing the chance of inappropriate financial guidance, confusion, or misuse in contexts where regulated or high-stakes advice should be handled carefully.

Vague Triggers

Medium

• Confidence: 95% confidence
• Finding: Examples like '风险评估' and '今日市场分析' are generic and overlap heavily with ordinary conversation, making accidental invocation plausible. In a financial domain, unintended activation is more concerning because the skill may generate authoritative-sounding trading analysis or strategy suggestions in response to vague prompts, creating user trust and compliance risks.

Natural-Language Policy Violations

Medium

• Confidence: 90% confidence
• Finding: The template is entirely written to operate in Chinese and does not provide any mechanism to honor the user's preferred language. In a multi-agent skill, this can cause instruction mismatch, reduced transparency, and user misunderstanding of financial analysis or risk disclosures, especially when outputs may influence investment decisions.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The skill explicitly states it will execute stop-loss/take-profit instructions and trading directives, but it provides no guardrails such as user confirmation, simulation-only defaults, broker/account scope checks, or warnings about real financial impact. In an investment-agent context, this creates a meaningful risk of unauthorized or unintended real trades that could directly affect user assets.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The workflow describes monitoring triggers and executing risk-control operations and trading instructions as part of normal behavior, again without documenting any safety interlocks or impact warnings. Because this is a risk-management agent embedded in a multi-agent quant trading skill, the surrounding context makes operational execution more plausible and therefore increases the chance of harmful asset-impacting actions.

Missing User Warnings

Medium

• Confidence: 88% confidence
• Finding: The template explicitly instructs monitoring social media discussions, sentiment, and investor group behavior, but it provides no privacy notice, consent boundary, or data-handling constraints. In an investment context, this can lead operators or downstream agents to collect, infer, and act on personal or sensitive behavioral data without clear minimization or compliance safeguards.

Natural-Language Policy Violations

Medium

• Confidence: 88% confidence
• Finding: The file is entirely in Chinese and contains no mechanism to detect or honor the user's preferred language, which can cause the agent to respond in an unexpected locale. In a financial analysis skill, this can reduce user comprehension of investment guidance, consent prompts, and risk disclosures, increasing the chance of misunderstanding or misuse.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: This section gives actionable investment decision guidance, including contrarian buy/sell timing, position management, and emotion-cycle-based signals, but does not clearly state that the material is educational only and not personalized financial advice. In the context of a multi-agent quantitative investing skill, users may reasonably treat these recommendations as operational trading guidance, increasing the risk of financial harm or regulatory/compliance issues.

Natural-Language Policy Violations

Medium

• Confidence: 87% confidence
• Finding: The guide requires every investment-analysis response to begin with a fixed Chinese disclaimer and forbids modifying the wording, with no language choice or user opt-in. This can override user language preferences and reduce transparency or accessibility for non-Chinese speakers, which is a real quality/safety issue in a user-facing skill, especially in a regulated financial context where users must clearly understand disclosures.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The monitoring specification explicitly requires logging input parameters and output results, but it does not require redaction, minimization, or exclusion of sensitive data. In an investment-analysis multi-agent system, those fields can contain API credentials, portfolio positions, trading instructions, user financial data, or proprietary strategy details, so broad logging materially increases the risk of data exposure through log access, retention, or downstream observability tools.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: This file gives concrete trading guidance, entry/exit rules, position sizing, and risk-management instructions, but does not include a clear user-facing warning that investing can result in substantial losses and that the material is educational rather than personalized financial advice. In the context of a 'Wall Street quant team' skill, users are especially likely to rely on this guidance for real-money decisions, increasing the chance of financial harm from overconfidence or misuse.

Missing User Warnings

High

• Confidence: 97% confidence
• Finding: The emergency risk-control flow states that stop-loss may be executed immediately within 60 seconds, but does not require prior user notification or confirmation even though it can directly affect holdings and assets. In the context of an investment skill, this creates a high-risk pathway for unauthorized liquidation, mistaken triggers, or abuse through manipulated signals, causing real financial loss.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The extractor calls zipf.extractall(output_path) on an arbitrary .skill/.zip archive without validating member paths. A crafted archive can use path traversal entries such as ../ or absolute paths to write files outside the intended extraction directory, potentially overwriting sensitive files or planting malicious content; in this skill context, users may reasonably handle untrusted packaged skills, which increases the danger.

Static analysis

No suspicious patterns detected.