Back to skill

Security audit

去AI味文本润色助手 · Su's StoicLab

Security checks for vulnerabilities and agentic risk

Overview

This is a text-editing skill that rewrites AI-sounding prose and does not request code execution, data access, persistence, or external connections.

Install this if you want a Chinese text-polishing skill focused on reducing AI-like style. Be aware it may activate for broad editing or review requests, so users who only want ordinary proofreading may need to specify that they do not want de-AI-style rewriting.

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
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 (3)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The description says to use the skill for tasks like article polishing, review checks, manuscript revision, and colloquial rewriting, then adds '编辑或审阅任何疑似AI生成的文本时使用本技能.' This mixes a specific purpose with broad common editing activities, without clear boundaries or exclusion cases, increasing the chance of unintended invocation during normal writing-help requests.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

Phrases like '润色', '审稿', and the scenario list '编辑公众号文章、审阅稿件、修改草稿、口语化改写' are generic editing activities, not uniquely tied to this skill's specialized function. Because no negative examples or stricter trigger constraints are provided, the skill could activate for many ordinary editorial tasks.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
86% confidence
Finding

L92 says to output only the revised text and provide no extra explanation or supplementation, but L93 then permits adding a separate "编辑说明" when special context needs explanation. This documentation creates conflicting intent about whether the skill ever produces explanatory output.

Content

No source excerpt is available for this finding.