Back to skill

Security audit

AU Modern Award Payroll Auditor

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed Australian payroll-audit prompt with no code or hidden access, but its legal and pay-rate outputs should be checked against current official sources.

Install this only if you want an Australian Modern Award payroll compliance helper. Treat its outputs as preliminary, avoid sharing unnecessary employee-identifying details, and verify all rates and legal conclusions with the Fair Work Ombudsman/Fair Work Commission or a qualified professional before making payroll, disclosure, or employee-communication decisions.

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
93% confidence
Finding

The trigger list is broad enough to activate on many ordinary workplace, payroll, or employment discussions, increasing the chance the agent enters this specialized skill when the user did not ask for award-compliance analysis. In a legal/payroll context, unintended invocation can steer conversations into incorrect regulatory framing, collect unnecessary sensitive employment details, or provide overconfident compliance guidance outside the skill's intended scope.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

At L441-L442, the guardrail says to never advise the user to take or avoid legal action and to direct them to the Fair Work Ombudsman or a lawyer instead. However, the error-handling workflow at L541-L542 goes beyond neutral referral by recommending when to seek a lawyer's advice as part of a remediation strategy, which creates a documentation-level contradiction about the intended boundaries of assistance.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The rule at L451 prohibits stating specific base-rate dollar amounts from memory and requires verification through official tools. Yet the examples at L565-L567 and L628-L631 calculate and present concrete or approximate rates using assumed base figures supplied or inferred in the example narrative, which undercuts the stated instruction not to provide memorized figures.

Content

No source excerpt is available for this finding.