Back to skill

Security audit

subclip-motion-graphics

Security checks for vulnerabilities and agentic risk

Overview

The verdict is benign because the scanners found no strong evidence of harmful behavior and assessed the one reported concern as low probability, though coverage was incomplete.

The overall verdict is benign, but not a guarantee of safety. Skillspector's analysis was partial: a static tool-misuse check hit a parser limit, so some evidence is incomplete. Consider whether the skill's reported preference for choosing a framework without asking fits how you want it to work before installing.

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
  • 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 · SKILL.md (reported line 22)May include surrounding context.

md
ader, a single badge or card entrance, or an exact inline SVG treatment. Never build a video's scenes, a demo, an explainer, a promo, or any multi-beat sequence from HTML/CSS items, even when each scene looks simple. Use HyperFrames for those: creators can click into a HyperFrames scene to edit its text and shapes directly, and HyperFrames scenes are easier for them to adjust later.

An explicit framework or exact template request always wins. When the creator does not name one, use HyperFrames without asking them to choose; "it would be shorter in HTML/CSS" is not a reason to switch engines for a scene. Preserve an existing layer's engine when editing it; do not migrate or replace it without a request. Do not place scripts in a native HTML/CSS Motion item.

An explicit request to adapt or use HyperFrames registry blocks routes to `$subclip-hyperframes-layer`. Preserve every named block slug: look up each one through the registry tools before authoring instead of replacing it with a ge