The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill is configured to interact with an external service at https://video-editor.ai-creator.top/. The use of the .top top-level domain is a potential risk indicator, as it is commonly utilized for hosting malicious content, and the domain does not belong to a recognized trusted organization.
[COMMAND_EXECUTION]: The skill executes a bundled Node.js script located at scripts/validate_edit_plan.mjs. More critically, it instructs the agent to discover and run "Agent command scripts" defined within a project's package.json file, which allows for the execution of arbitrary local commands if a repository contains a malicious configuration.
[DATA_EXFILTRATION]: By directing the agent to browse the external ai-creator.top site while processing local media and project files, there is a risk that sensitive project metadata or media information could be exfiltrated to the external domain during interaction.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it ingests untrusted media metadata and transcripts. * Ingestion points: Local media assets, .timeline project archives, and repository-specific package.json files. * Boundary markers: None provided; the instructions do not guide the agent to isolate or treat external media data as untrusted. * Capability inventory: Includes file system access (read/write), local script execution, and automated browser interactions. * Sanitization: No evidence of sanitization or validation of external content (such as transcription text or media metadata) before it is processed by the agent.

edit-timeline-studio