ClawHub
Back to skill

Security audit

Ai Video

Security checks for vulnerabilities and agentic risk

Overview

This is a non-executable AI-video routing and briefing guide with clear consent and disclosure guardrails and no evidence of hidden execution or data theft.

Before installing, treat it as planning guidance for AI video workflows. Verify tool availability and platform disclosure rules, use only consented likenesses or licensed talent, and keep a human review step before generation or publishing.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · references/tools-scope-and-disclosure.md (reported line 34)May include surrounding context.

md
- **AI disclosure is mandatory:** EU AI Act transparency; **TikTok auto-discloses**; **YouTube
  Altered-Content** label; disclose in caption and/or on-screen. Never strip a disclosure to "look real."
- **No non-consensual likeness/voice.** Never synthesize a real person (celebrity, competitor's
  staff, anyone) without consent — that's impersonation/rights infringement. Use consented avatars/
  digital twins or licensed talent.
- **Rights caution:** the generative-video field has live litigation (e.g. the Hailuo/MiniMax
  copyright suit) — review rights, resemblance, and usage terms before commercial publishing.

Static analysis

No suspicious patterns detected.