Security audit

Marketing Orchestrator Adarsh

Security checks for vulnerabilities and agentic risk

Overview

This skill is a small, disclosed marketing-audit orchestrator that coordinates other collector skills and does not include executable code, persistence, or hidden behavior.

Before installing, confirm you are comfortable with a workflow that may call separate marketing collector skills and use configured API keys for services such as social, ads, SEO, competitor, and website-audit data. The main improvement would be clearer failure handling and a more explicit privacy notice, but the reviewed artifact itself is coherent and limited.

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

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (2)

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: This markdown file describes orchestrating multiple collector agents, aggregating raw data, and generating a final report, and later notes that the framework should handle API keys and environment variables. However, it does not warn users that running the skill may access external services, process collected marketing data, or rely on credentials, which are behaviors that can affect privacy and system integrity.

Intent-Code Divergence

Medium

Confidence: 98% confidence
Finding: The documented behavior says individual collector failures should not block orchestration, but the pseudocode directly awaits each sub-skill without try/catch or fallback handling. In practice, any collector error would terminate the workflow and prevent report generation, creating a denial-of-service style reliability failure and potentially causing partial data loss for the audit run.

Static analysis

No suspicious patterns detected.