Back to skill

Security audit

Ai Workflow Master

Security checks for vulnerabilities and agentic risk

Overview

This skill is not malicious, but it should be reviewed because it can activate from very generic requests and may process files, URLs, customer data, or create tickets without clear confirmation rules.

Review this skill before installing. It is Markdown-only and shows no hidden code, but users should narrow its activation phrases and require explicit confirmation before file or URL ingestion, customer-data processing, external scraping, publishing-related actions, or ticket creation.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • 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 (3)

Vague Triggers

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill metadata advertises very broad trigger phrases such as '做这个', '帮我搞定', and '自动化', which are common in normal conversation and not tightly bound to this skill’s domain. This creates a real risk of unintended activation, causing the assistant to enter a multi-step workflow mode when the user did not explicitly request this specific skill.

Content

No source excerpt is available for this finding.

Vague Triggers

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The activation section further expands ambiguous triggers like '帮我做', '一键生成', '快速完成', and '提高效率' without any scoping or guardrails. In context, this is more dangerous because the skill chains multiple capabilities including data handling and automated workflow execution, so accidental activation could lead to unintended processing of user content or overbroad task execution.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The workflows include data collection from URLs, file reading, public-data scraping, customer-issue handling, emotion analysis, and automatic ticket creation, but there are no clear user-facing privacy notices, consent checkpoints, or data-minimization constraints. This is risky because users may provide sensitive files, links, or customer content without understanding that the skill may ingest, analyze, and persist that data across workflow steps.

Content

No source excerpt is available for this finding.