Security audit

内容创业实操教练

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly a business-coaching package, but it adds off-platform activation and persistent local learning behavior that users should review before installing.

Install only if you are comfortable with a manual author-registration step, a unique local installation ID, and persistent local notes about your usage. Before using it with sensitive business plans, disable or review the learning log and avoid sending more identifying information to the author than necessary. No automatic network exfiltration, destructive behavior, or remote-code retrieval was found in the inspected artifacts.

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
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection

Findings (8)

Tp4

High

Category: MCP Tool Poisoning
Confidence: 99% confidence
Finding: The declared purpose is business coaching, but the skill embeds a licensing/activation workflow that generates a unique installation identifier, stores it locally, and instructs the user or agent to contact the author for registration before full functionality is available. This is a material behavior mismatch that can mislead users and operators about what the skill actually does, and it introduces tracking and off-platform authorization dependencies unrelated to the advertised function. The danger is heightened by instructions telling the AI to generate the identifier itself when Python is unavailable, which normalizes hidden state creation and gating without clear disclosure in the metadata.
Content: No source excerpt is available for this finding.

Undeclared Tool Scope

Medium

Category: MCP Least Privilege
Confidence: 93% confidence
Finding: The skill directs the host agent to run local scripts and write files, but it does not declare any explicit tool or permission scope. That creates an authorization gap: an agent may perform file reads/writes and code execution behavior that the skill metadata does not transparently disclose or constrain. In this context, the risk is increased because the skill also persists activation state and learning logs on disk, so undeclared file access is part of normal operation rather than incidental.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 91% confidence
Finding: The description says to use the skill whenever the user asks about a wide range of topics including content entrepreneurship, side hustles, personal branding, private-domain customer acquisition, monetization, and starting a one-person company. This spans many common advice scenarios without clear exclusion conditions or negative examples, making unintended invocation more likely.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The skill instructs persistent storage of user/task-derived feedback and usage data into a local memory log, but it does not clearly warn the user at the moment of collection or obtain consent for retention. Even if storage is local, these notes can contain sensitive business plans, performance metrics, or proprietary ideas, creating privacy and data-handling risk. The surrounding 'self-learning' framing makes this more dangerous because persistence is normalized as an automatic post-task step.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 94% confidence
Finding: The file explicitly instructs ongoing logging of 'new experiences' after each use and weekly consolidation, but provides no notice about persistence, storage location, retention limits, or restrictions on what may be recorded. In an agent skill context, this can lead to silent accumulation of user-derived or session-derived data, including potentially sensitive business, personal, or operational details, without informed consent or minimization.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium

Category: Not specified by scanner
Confidence: 96% confidence
Finding: The script adds a manual activation and authorization mechanism that is unrelated to the stated purpose of a content/coaching skill and creates an unnecessary control channel between the end user and the author. Even though it does not exfiltrate data automatically, it conditions full functionality on sharing a locally generated identifier with the author, which introduces tracking, gating, and potential misuse outside the skill’s declared function.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

Category: Not specified by scanner
Confidence: 86% confidence
Finding: SQP-3 applies to all file types and includes language or locale policy violations. The title and operating instruction are written as mandatory Chinese-language guidance, with no indication that users may choose another language or that the skill is intentionally region-specific.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Low

Category: Not specified by scanner
Confidence: 90% confidence
Finding: The documentation says the identifier is only generated and stored locally and not uploaded automatically, but the script explicitly instructs the user to send that identifier to the author for registration. This discrepancy is misleading and can cause users to disclose installation-linked data under a false sense that it remains purely local.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.