Security audit

Ads Creative Performance

Security checks for vulnerabilities and agentic risk

Overview

This is a markdown-only advertising guidance skill with some naming and trigger-scope mismatch, but it does not install code, access credentials, run commands, persist, or mutate accounts.

Install only if you want advertising creative planning and testing guidance, not a rigorous post-launch metrics analyzer. Users should provide actual performance data when asking for performance judgments and should treat any budget or campaign advice as advisory unless independently validated.

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 (3)

Description-Behavior Mismatch

Medium

Confidence: 96% confidence
Finding: The skill manifest promises post-launch creative performance analysis, but the workflow and outputs are centered on generating new creatives, hooks, and test plans. This semantic mismatch can cause an orchestrator or user to invoke the skill in the wrong context, leading to unsafe automation decisions, misleading outputs, and improper handling of performance-analysis tasks that may require different data, controls, and escalation paths.

Intent-Code Divergence

Medium

Confidence: 94% confidence
Finding: The body of the skill repeatedly documents pre-launch creative generation, adaptation, and scoring behavior while branding itself as post-launch performance analysis. These contradictions increase the chance that an agent system routes sensitive business decisions to a skill that performs a different function than advertised, degrading reliability and potentially causing harmful optimization or budget recommendations based on absent or assumed evidence.

Vague Triggers

Medium

Confidence: 95% confidence
Finding: The trigger list includes broad business terms like revenue, profit, campaign, and growth, which are likely to match many unrelated requests. Over-broad activation raises the risk of unintended tool invocation, context confusion, and leakage of irrelevant or low-confidence advice into workflows that did not request advertising creative analysis.

Static analysis

No suspicious patterns detected.