Security audit

AI KEY·口播剪辑

Security checks for vulnerabilities and agentic risk

Overview

This skill provides a clearly scoped video-editing workflow with user checkpoints and proportionate access to media and editing notes.

Check ChatCut's processing and retention before importing private footage; use local transcription when appropriate. Keep source backups and limit cleanup to project copies. Expect editing notes to persist in your configured vault. Companion skills and external services are outside this package review.

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

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
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (3)

The footage-import workflow lacks a privacy warning about uploading user media to ChatCut.

Medium

Category: Output Handling
Confidence: 88% confidence
Finding: The workflow permits importing footage from a supplied local path into ChatCut without first explaining whether the media or transcript will leave the device. If the selected integration uploads content, private footage, client information, or financial details could reach an external service before the workflow's later redaction steps; authorization to edit a local file does not necessarily authorize that disclosure.
Content: No source excerpt is available for this finding.

The workflow instructs deletion of unmasked originals without warning about potential loss of source media.

Medium

Category: Output Handling
Confidence: 87% confidence
Finding: The instruction to delete unmasked originals after use is an actionable deletion step, not merely a tool description. Although it explains the privacy motivation, it does not warn that deleting the only source copy could permanently lose user media or prevent later re-editing.
Content: No source excerpt is available for this finding.

The prompt template mandates English camera terminology without user opt-in.

Low

Category: Output Handling
Confidence: 90% confidence
Finding: The instruction requires shot types and camera movements to be written using their English terms. This explicitly constrains part of the generated output to a particular language without offering a user choice or explaining a model compatibility requirement.
Content: No source excerpt is available for this finding.