ClawHub
Back to skill

Security audit

AdMakeAI

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed AdMakeAI API helper whose sensitive actions are aligned with creating and managing ads, with clear confirmation guidance for credit-spending and Meta write actions.

Install this only if you intend to let the agent use your AdMakeAI account for ad workflows. Keep the API key in the environment, review project/account choices carefully, and require explicit confirmation before generations, uploads, batch runs, or any Meta draft/publish action that could spend credits or create advertising objects.

SkillSpector

Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (2)

Vague Triggers

Medium
Confidence
93% confidence
Finding
The trigger description is extremely broad and includes many common verbs and ad-related terms, which can cause the skill to activate in situations where the user did not specifically intend to use AdMakeAI. Because this skill has network access and can perform credit-spending or external write actions, over-invocation increases the chance of unintended API calls, unnecessary data exposure, or prompting users into workflows they did not request.

Unrestricted Tool Access

Medium
Category
Excessive Agency
Content
`Authorization: Bearer <key>` also works. Never echo the key back to the user or paste it into a file.

## How to call

Every tool has one canonical tRPC dotted path — `adGeneration.create`. Two transports expose the same allowlist:
Confidence
75% confidence
Finding
The skill documentation presents a broad, effectively unrestricted procedure surface ('Every tool has one canonical tRPC dotted path') over direct HTTP/MCP access, while the allowed tools include general network and shell capabilities. Even though the document gives policy guidance about confirmation for dangerous actions, there is no technical enforcement in the skill file itself to prevent use of sensitive or credit-spending endpoints once the skill is invoked.

Static analysis

No suspicious patterns detected.