Security audit

private-equity

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent private-equity workflow, but it can write sensitive deal materials to Notion when connected without a clear per-use confirmation step.

Review before installing if your PE/VC materials are confidential. Keep Notion disconnected or require explicit approval before any export, redact sensitive deal terms and company data before sharing with external tools, and have legal, tax, and investment professionals validate outputs before acting on them.

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

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

Findings (13)

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 86% confidence
Finding: Several trigger terms in the routing table are generic business phrases such as '投委会', '回报率', and 'IPO退出', which may appear in ordinary discussion outside the intended skill context. Because this skill handles investment decisions, diligence, and legal-commercial review, ambiguous triggers raise the chance of accidental activation, misrouting, or inappropriate processing of confidential transaction materials.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The README says users can trigger the skill through natural conversation and that matching is automatic, but it does not define narrow routing boundaries or disambiguation rules. In a multi-skill environment, broad natural-language activation can cause this finance-oriented skill to run on loosely related requests, potentially exposing sensitive deal, legal, or financial data to the wrong workflow or producing unintended analyses.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 94% confidence
Finding: The description embeds many short trigger phrases such as “尽调清单”, “due diligence”, “投委会”, and similar generic business terms. Broad triggers increase the chance that unrelated finance or legal-assistance conversations activate this skill unexpectedly, causing the agent to follow private-equity workflows when the user intended a narrower or different task.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 92% confidence
Finding: The routing logic activates a subskill whenever user input contains keywords or loosely matching scenarios, then selects the 'highest match' without a precise scoring boundary. This ambiguity can misroute users into the wrong subskill, leading to inappropriate document handling, incorrect financial/legal outputs, or unintended use of downstream templates.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 90% confidence
Finding: This file presents the skill name and instructions in a single forced language, and no nearby text indicates that users may choose another language or that the skill is intentionally limited to a Chinese-only audience. Under the policy rule, forcing a specific language without user opt-in is a natural-language policy concern.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 94% confidence
Finding: The skill instructs writing due-diligence checklists into Notion, which can include highly sensitive company, legal, financial, and compliance information, but it provides no warning, consent checkpoint, or data-minimization guidance before using an external SaaS platform. In a private-equity due-diligence context, this increases the risk of unintended disclosure, over-sharing, or violating confidentiality obligations to portfolio targets and counterparties.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 94% confidence
Finding: The skill explicitly directs writing the exit analysis into Notion, which can contain sensitive investment, fund, valuation, cap table, and transaction-term data, but it does not warn that this information may be transmitted to an external third-party service. In a private-equity context, these materials are often highly confidential and subject to contractual, regulatory, or fiduciary restrictions, so silent exfiltration to a connector materially increases data-leak risk.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The entire skill specification, required output template, and operational instructions are written to produce Chinese-language investment committee materials, but the file does not offer a language preference or opt-in. Under the stated policy, forcing a specific language without user choice is a natural-language policy violation unless clearly justified as region-specific.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 90% confidence
Finding: The title and the entire skill description are written exclusively in Chinese, with no indication that users may choose another language or that the skill is limited to a China-specific audience. Under the stated policy, forcing a specific language without user opt-in is a natural-language policy violation.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 89% confidence
Finding: This markdown skill specifies all instructions, inputs, and outputs exclusively in Chinese, which imposes a language constraint on users. The file does not provide an opt-in language choice or explain that the skill is intentionally limited to a Chinese-language or China-specific audience.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

Low

Category: Not specified by scanner
Confidence: 78% confidence
Finding: The documentation states at L003 that the skill can be used completely without connectors, framing operation as manual-input based. Later, L228-L232 introduces a Notion connector that writes project screening memos into an external knowledge base, which is additional behavior beyond the earlier self-contained description even if marked optional.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

Category: Not specified by scanner
Confidence: 74% confidence
Finding: The skill's title and all user-facing instructions are in Chinese, and there is no indication that the user may choose another language or that the skill is intentionally limited to a Chinese-language context. Under the stated policy, forcing a specific language without opt-in can be a natural-language policy violation.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Low

Category: Not specified by scanner
Confidence: 82% confidence
Finding: L003 明确表示本技能的全部功能都基于用户提供的文件和文本、无需连接器即可完整使用；但 L165 与 L173 又写明存在 Notion 连接失败回退以及将报告写入 Notion 知识库的能力。这不是单纯的信息补充，而是对“所有功能”边界的表述产生了直接冲突。
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.