ClawHub
Back to skill

Security audit

Ad copy frameworks

Security checks for vulnerabilities and agentic risk

Overview

The skill stays within writing and checking ad copy, with a routine dependency risk from its unpinned counting tool.

Before using the counting command, choose a reviewed local counter or approve and pin the external package version. The npx command may download and execute third-party software. The skill itself does not authorize publishing ads or accessing advertising accounts.

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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • 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)

MCP server referenced without pinned version: 'npx social-media-specs'.

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

The command invokes an unpinned npm CLI package, not an MCP server, so the finding's classification is inaccurate but its supply-chain concern is valid. If the package is not available locally, npx can retrieve and execute a registry version; a compromised release could run code with the agent's permissions. Character counting is relevant to this skill, but it does not require executing an unreviewed dependency.

Content

No source excerpt is available for this finding.

The prescribed character-counting command can download and execute third-party code without a user warning.

Medium
Category
Output Handling
Confidence
94% confidence
Finding

The skill directs the agent to use npx for a simple counting task without explaining that this may download and execute third-party code. Although npx may request installation confirmation depending on its configuration, that is not a reliable security boundary for automated agents. This overlaps with RP1 and creates unnecessary execution exposure, rather than demonstrating malicious intent.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.