Security audit

Agent评估设计顾问

Security checks for vulnerabilities and agentic risk

Overview

This is a documentation-only skill for advising whether a workflow fits an AI agent, with no code execution, persistence, or hidden data access.

Install this if you want Chinese-language guidance for deciding whether a workflow should use an AI agent. Be aware it is intentionally advisory only, avoids specific model names and code implementation, and may need language-preference clarification if used in non-Chinese conversations.

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
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (4)

Natural-Language Policy Violations

High

Confidence: 95% confidence
Finding: The skill content is written entirely in Chinese and implicitly constrains behavior and output format without any indication that language should follow the user's preference. In a multilingual assistant, this can cause unwanted language switching, reduce usability, and interfere with instruction-following if the user interacts in another language.

Vague Triggers

Medium

Confidence: 91% confidence
Finding: The trigger condition is broad enough that the skill may activate on loosely related workflow-design questions, causing the assistant to apply this skill outside its intended scope. While this is not directly exploitable for code execution or data exfiltration, over-broad activation can override better-matched instructions and degrade safety, accuracy, or user intent handling.

Vague Triggers

Medium

Confidence: 93% confidence
Finding: The activation guidance says to activate when a user describes a workflow and wants to judge agent suitability, but it does not clearly state when the skill must not activate. This ambiguity can lead to inappropriate invocation on partial matches, increasing the chance of misrouting the conversation and applying restrictive instructions in contexts where they do not fit.

Natural-Language Policy Violations

Low

Confidence: 94% confidence
Finding: This markdown file is entirely written in Chinese and does not indicate that the language is optional, user-selected, or required for a specific regional/compliance context. Under the language/locale policy rule, forcing a specific language without user opt-in is a natural-language policy concern.

Static analysis

No suspicious patterns detected.