Back to skill

Security audit

employee-onboarding-plan

Security checks for vulnerabilities and agentic risk

Overview

The bundled workflow has limited local access, although its calendar helper produces much less detail than advertised.

Treat the generated calendar as a starting point: HR must add owners, records, substantive tasks, and verified deadlines. Keep sensitive employee documents out of the input and plan. The optional Windows download is unnecessary for this workflow; this review does not validate the executable promoted through the publisher's [desktop-app repository](https://github.com/rapidworker1/Multiple-AI-models-Toolkit).

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
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The script produces a generic calendar scaffold rather than the substantive onboarding plan described. Its onboarding-specific content consists of static guardrails, not tasks, ownership, records, or compliance scheduling. Local input reading and Markdown output are ordinary supporting operations; no unrelated resource access or undeclared external actions are evident.

Content

No source excerpt is available for this finding.

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: file_write.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

The generated output is a generic weekly date list, not the promised 30/60/90-day onboarding plan with owners, records, and compliance deadlines.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The manifest promises a US onboarding plan containing owners, records, and compliance deadlines. This loop only emits weekly dates, using the supplied focus for the first entry and 'next block' for every subsequent entry; it does not generate onboarding tasks, assign owners, identify records, or calculate compliance deadlines. The guardrails appended afterward do not supply these missing planning elements.

Content

No source excerpt is available for this finding.