Security audit

Ads Landing Page Audit

Security checks for vulnerabilities and agentic risk

Overview

This is a markdown-only advertising audit skill with some overbroad marketing-operations guidance, but no code, persistence, credential handling, or automatic account actions.

Install only if you want an ads operations and landing-page audit helper, not a narrowly limited page checker. Review its recommendations before acting on budget, bidding, containment, or escalation advice, since the skill does not itself verify live ad account data or perform account changes.

A.I.G

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

SkillSpector

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 (4)

Description-Behavior Mismatch

High

Confidence: 97% confidence
Finding: The skill metadata and title present a narrow landing-page audit capability, but the body expands into broad ads operations, incident response, and account-level campaign guidance. This scope deception can cause incorrect routing, over-privileged invocation, or reliance on the skill for decisions outside its declared domain, increasing the chance of unsafe or unauthorized business actions.

Intent-Code Divergence

Medium

Confidence: 90% confidence
Finding: The skill frames itself as a landing-page auditor, but its workflow and outputs produce incident-response artifacts like escalation tickets, anomaly handling, and platform-ops mitigation actions. This mismatch is dangerous because downstream systems or users may trust the skill for a narrow audit task while it actually steers operational responses and campaign decisions.

Description-Behavior Mismatch

Medium

Confidence: 94% confidence
Finding: The trigger section broadens activation from landing-page auditing to generic ad execution, growth, revenue, budget, and platform optimization requests. Overbroad triggers increase the likelihood that the skill is invoked for unrelated or higher-risk business guidance, creating prompt-routing confusion and unintended authority expansion.

Vague Triggers

Medium

Confidence: 95% confidence
Finding: The trigger conditions are broad enough to match many ordinary marketing or business questions, which can cause accidental invocation and inappropriate use of specialized guidance. In agent systems, overly permissive trigger logic is a security and safety concern because it expands the skill's effective reach beyond its reviewed purpose.

Static analysis

No suspicious patterns detected.