Security audit

Video Creator

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly aligned with video creation, but it needs review because its recipes can overwrite or delete local files and install an unpinned dependency without clear safeguards.

Install only if you are comfortable with a video skill that runs local ffmpeg/Python commands. Use it in a dedicated project folder, review generated commands before execution, avoid running recipes in shared directories, be careful with overwrites, install Python dependencies in an isolated environment, and use yt-dlp only for media you own or are authorized to download and remix.

A.I.G

Vulnerability Patterns

Insecure DependenciesIntroduces malicious components through unsafe dependency sources
Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
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

Findings (2)

T08 · Insecure Dependencies

Warning

Location: references/ffmpeg-recipes.md:430
Finding: Unpinned Third-Party Dependency Installation
Content: View full analysis
Remediation: View remediation

T09 · Insecure Skill Coding Practices

Warning

Location: references/ffmpeg-recipes.md:364
Finding: Predictable Temporary Files and Unsafe Working-Directory Cleanup
Content: View full analysis
Remediation: View remediation

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
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
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (4)

Vague Triggers

High

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The activation text says to use this skill 'ALWAYS' for a very broad set of video-related terms, causing the skill to trigger in many contexts beyond simple video assembly. Over-broad invocation increases the chance the agent will enter a code-executing, file-writing workflow unnecessarily, which can lead to unintended actions or misuse in ambiguous requests.
Content: No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium

Category: Not specified by scanner
Confidence: 90% confidence
Finding: The skill explicitly lists yt-dlp as an optional tool to download YouTube videos 'for reference/remix,' which expands the skill from user-provided asset processing into third-party content acquisition. That creates legal/compliance risk and can enable unauthorized downloading or reuse of external media not necessary for the stated purpose of creating videos from assets.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: The response format directs the agent to 'Execute and deliver: Run the script and show the result' without any safety note about filesystem writes, overwriting outputs, using external tools, or confirming execution. In context, this skill routinely generates ffmpeg/Python commands that create or modify local files, so silent execution can cause unintended changes or destructive overwrites.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 94% confidence
Finding: The silence-removal section includes commands and script logic that forcibly overwrite outputs (-y) and delete temporary files with wildcard cleanup (rm _seg_*.ts _concat_list.txt) without prominent warnings or safer guards. In an agent skill context, users may copy or the agent may adapt these commands into working directories containing similarly named files, causing unintended data loss or destructive side effects beyond the immediate task.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.