Serenity-framework/白毛股神的分析框架

• Outcome: Review
• Latest audit: Sep 16, 2026 · 9:39 AM UTC
• Version: 1.0.4

Overview

This is a stock-analysis skill with no executable code, but it gives concrete trading and position-sizing guidance and can trigger external financial lookups without enough user-facing safeguards.

Review this carefully before installing if you might use it for real money decisions. Treat its outputs as research prompts only, not personalized investment advice, and require explicit consent before web or finance-tool lookups. Be especially cautious with position sizes, margin, day trading, and persona-style recommendations.

A.I.G

Vulnerability Patterns

• Skill Instruction Hijacking: Alters the agent's session goals or safety constraints when the skill loads
• Agent Memory Poisoning: Writes attacker-controlled rules into memory that affect later sessions
• Remote Payload Retrieval and Execution: Fetches external code whose behavior can change after review
• Embedded Malicious Code: Ships malicious scripts inside the skill and executes them locally
• Unauthorized Access and Privilege Escalation: Obtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (17)

Missing User Warnings

Medium

• Confidence: 96% confidence
• Finding: The skill explicitly offers position sizing, allocation percentages, stop-thesis exits, and leverage-oriented guidance, but does not present a clear, prominent user-facing warning that the output is not financial advice and may cause real financial loss. In an investment-analysis skill, this omission is materially risky because users may treat the recommendations as actionable personalized advice rather than educational analysis.

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: The skill instructs automatic online retrieval of financial and web data when certain triggers are met, but does not clearly warn users that external services may receive their queries and that retrieved information may be inaccurate, stale, or privacy-sensitive. This is more dangerous in context because the skill analyzes user-supplied holdings, tickers, and theses, which can reveal sensitive financial interests to third-party tools without meaningful notice or consent.

Intent-Code Divergence

Medium

• Confidence: 98% confidence
• Finding: The English and Chinese sections describe materially different source corpora and date ranges used to derive the framework (3,704 tweets from 2025-07-02 to 2026-06-25 versus 839 tweets from 2025-12-26 to 2026-06-24). In an investment-analysis skill, this undermines provenance, reproducibility, and user trust; a user could rely on a methodology whose claimed evidence base changes depending on language, which can mislead downstream financial decisions.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: This section gives specific, actionable trading instructions on allocation sizing, averaging up, trimming, and stop-loss thresholds without any warning that the content is not personalized financial advice and may cause significant losses. Because the skill is explicitly framed as a stock-analysis framework, users may over-trust these heuristics and apply risky position-management rules unsafely.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The Chinese section repeats the same concrete trading instructions, including leverage-like conviction sizing behavior and wide stop-loss thresholds, but still provides no user-facing warning about financial risk or that outcomes can vary materially. This broadens exposure because bilingual users receive equivalent actionable guidance without safeguards in either language.

Missing User Warnings

Medium

• Confidence: 97% confidence
• Finding: This file gives actionable trading instructions, including day-trading tactics, profit targets, loss limits, and language about using an information advantage to 'front-run market reaction,' but it provides no warning about financial risk, suitability, or the possibility of substantial loss. In a skill explicitly designed to guide stock trading decisions, that omission materially increases the chance users will treat the content as safe operational advice and incur significant financial harm.

Description-Behavior Mismatch

Medium

• Confidence: 95% confidence
• Finding: The reference file does not merely analyze an investing style; it includes a reusable prompt that instructs the agent to imitate a real social-media persona and generate recommendation-style investment output. That materially expands the skill from research support into persuasive persona emulation and advice generation, increasing the chance of misleading users, misattribution, and unsafe financial guidance.

Context-Inappropriate Capability

Medium

• Confidence: 94% confidence
• Finding: The document provides actionable guidance for audience-tailored persuasion, including how to speak differently to English and Chinese audiences, when to use intimacy cues, and how to frame conviction statements. In skill context, this can steer the model toward manipulative, credibility-borrowing behavior rather than objective analysis, especially because it is tied to a named influencer persona.

Missing User Warnings

Medium

• Confidence: 97% confidence
• Finding: The embedded 'signal radar' and position-sizing rules give concrete, actionable trading heuristics, including triggers for entry, position sizes, DCA cadence, and drawdown handling. Because the skill is for stock analysis, this context makes the issue more dangerous: users could treat the output as investment advice, leading to financial harm, especially when rules are presented with unwarranted specificity and authority.

Natural-Language Policy Violations

Low

• Confidence: 77% confidence
• Finding: The file hardcodes two specific languages for the skill content and does not indicate that language selection is optional or user-driven. Under the stated policy, fixed language behavior without opt-in can qualify as a locale/language policy concern unless clearly justified.

Natural-Language Policy Violations

Low

• Confidence: 87% confidence
• Finding: This markdown file includes parallel English and Chinese sections as a fixed presentation format, which can constitute a language/locale policy issue when no user preference or opt-in is offered. The policy explicitly allows locale constraints only when they are optional or clearly justified, which is not stated here.

Natural-Language Policy Violations

Low

• Confidence: 88% confidence
• Finding: The document is structured to present both English and Chinese sections as the default format, but it does not state that language selection is optional or user-driven. Under the policy, a skill should not impose a specific language or locale without opt-in unless the constraint is clearly justified.

Natural-Language Policy Violations

Low

• Confidence: 78% confidence
• Finding: The file is structured as fixed English and Chinese sections, indicating a predetermined language/locale presentation. Under the policy, language constraints should either offer user choice or be clearly justified as region-specific; neither is stated here.

Natural-Language Policy Violations

Low

• Confidence: 85% confidence
• Finding: This markdown file includes a full Chinese-language section in addition to English, but it does not state whether language should be selected based on user preference or context. Under the policy, locale behavior should be opt-in or clearly justified when a specific language presentation is imposed.

Natural-Language Policy Violations

Low

• Confidence: 78% confidence
• Finding: The document is written in fixed English and Chinese sections and does not indicate that language selection is based on user preference or opt-in. Under the natural-language policy rule, imposing a specific language or locale without user choice can be a policy concern unless clearly justified.

Natural-Language Policy Violations

Low

• Confidence: 88% confidence
• Finding: This markdown file presents the skill template entirely in Chinese and frames its usage instructions accordingly, with no indication that users may choose another language. Under the policy rule for natural-language violations, forcing a specific language without user opt-in can be a locale/language policy issue.

Natural-Language Policy Violations

Low

• Confidence: 94% confidence
• Finding: The prompt specifies '中文模式（激活条件：对象含中文读者或 A 股标的）', which hard-codes a language switch based on audience or topic rather than user preference. This is a natural-language policy concern because it imposes a locale/language behavior without explicitly offering the user a choice.

Static analysis

No suspicious patterns detected.