Back to skill

Security audit

ResearchMate

Security checks for vulnerabilities and agentic risk

Overview

This skill is a Chinese-language research helper that uses public web sources and creates research output files, with no evidence of hidden, destructive, or persistent behavior.

Install this if you want a Chinese-language assistant to gather public research sources and generate local material files. Expect it to perform web searches, read public pages, and write Markdown/CSV-style outputs; avoid using it for private, paid, login-only, or sensitive internal information.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (5)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The skill documentation presents the assistant, prompts, invocation phrases, and examples entirely in Chinese, including the interactive questions the user is expected to answer. There is no indication that users may choose another language or that the Chinese-only behavior is a necessary region-specific constraint.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The README markets one-shot automatic collection and export, but does not clearly warn that the skill may access external sources and create local files as part of execution. In agent environments, missing disclosure undermines informed consent and can lead to surprising data egress, network activity, or filesystem changes from a simple chat request.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger examples are broad enough to match common writing and research requests, which can cause the skill to activate in situations where the user did not clearly consent to automated web collection and report generation. In this skill's context, that increases the chance of unexpected network access, data aggregation, and file export during ordinary conversation-driven use.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The trigger description is broad enough to overlap with many ordinary writing-help requests, which can cause this skill to activate when a simpler or different skill would be more appropriate. Overbroad activation can lead to unnecessary tool use, unexpected web retrieval, and inadvertent expansion of the skill's authority beyond the user's actual intent.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The need-based triggers like '素材不够' or '文章缺乏数据支撑' are ambiguous and rely on loose interpretation rather than precise activation boundaries. This increases the chance of accidental invocation, causing the agent to collect external material or create files in situations where the user may only want lightweight assistance.

Content

No source excerpt is available for this finding.