Back to skill

Security audit

AI Creative Studio

Security checks for vulnerabilities and agentic risk

Overview

This skill generates local marketing drafts without hidden account access, publishing, tracking, or background activity.

Review all drafts before publishing: the generator can insert unsupported claims such as trial terms, shipping offers, and ratings. Choose a separate output folder, confirm privacy and consent arrangements before implementing tracking or retargeting, and explicitly authorize any publishing or advertising spend.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Anti-Refusal Statement

High
Category
Anti-Refusal
Confidence
85% confidence
Finding

Skill instructs the agent to omit warnings, disclaimers, or ethical commentary. Stripping safety caveats hides risk from the user and is a common jailbreak preamble.

Content

Scanner excerpt · references/campaign-playbook.md (reported line 43)May include surrounding context.

md
- Change one variable per test: hook, offer, or creative format - never two.
- Minimum 3 variations per ad set.
- Do not judge a creative before ~1000 impressions or ~20 clicks.
- Kill winners are rare - kill the bottom third, scale the top third.
- Keep a "loss log": what you tested, what it cost, what happened.

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

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 launch workflow enables tracking and audience retargeting without a privacy warning.

Medium
Category
Output Handling
Confidence
91% confidence
Finding

The workflow directs the agent to launch a page with tracking enabled and retarget people based on video-viewing behavior. It does not warn about collecting or using visitor and viewer activity for advertising; the later email privacy-policy checklist does not address this tracking and retargeting.

Content

No source excerpt is available for this finding.