Security audit

Auto Expert Team

Security checks for vulnerabilities and agentic risk

Overview

The skill is a Markdown-only package with no code execution or data exfiltration found, but its stated “negative example” purpose conflicts with broad operational instructions that can unexpectedly turn many tasks into expert-team simulations.

Install only if you deliberately want a Chinese-language expert-team orchestration framework. Review its broad triggers and consider narrowing them, especially before using it for sensitive work involving finance, legal, invoices, social posting, or automated workflows.

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

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

Findings (6)

Description-Behavior Mismatch

Medium

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The skill presents itself as a 'negative example' arguing against expert-team simulation, but then supplies a complete, actionable framework for building and operating such teams. This mismatch can mislead reviewers, users, or upstream systems about the skill’s true behavior, causing unintended activation and execution of more powerful orchestration logic than the description suggests.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The trigger list includes broad, common phrases such as '团队完成任务' and '团队协作完成任务', which can match many ordinary user requests. Overbroad triggers increase the chance the skill activates unexpectedly, injecting a large orchestration framework into unrelated tasks and altering model behavior without clear user intent.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The document claims AI should directly solve tasks without role simulation, yet instructs the model to analyze tasks, generate expert roles, produce role prompts, design collaboration, and execute through a simulated team. This internal contradiction is dangerous because it obscures the effective instruction hierarchy and can cause policy evasion-by-framing, where risky orchestration is hidden behind a supposedly critical or educational wrapper.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 91% confidence
Finding: Stating that the skill applies to 'any task requiring multi-role collaboration' creates an ambiguous activation boundary with no clear exclusion criteria. In context, this makes the skill more dangerous because it is capable of restructuring broad classes of normal tasks into a role-simulation workflow, potentially overriding simpler or safer handling paths.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium

Category: Not specified by scanner
Confidence: 98% confidence
Finding: The manifest frames the skill as a critique of multi-role expert-team simulation and says AI should directly face tasks instead of simulating human teams. This file, however, is entirely dedicated to defining and advertising hundreds of experts and dozens of expert teams, which operationally supports the very team-simulation model the manifest argues against.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: This markdown file is entirely presented in Chinese and the title explicitly identifies it as a Chinese expert catalog, but it does not state that the skill is China-specific or offer any language/locale opt-in. Under the policy rule, forcing a specific language without user choice is a natural-language policy violation.
Content: No source excerpt is available for this finding.