The Agent Skills Directory

[COMMAND_EXECUTION]: The skill directs the agent to execute various shell commands through the Remotion CLI for rendering, previewing, and probing video files. Additionally, it provides a PowerShell validation script (scripts/validate-skill.ps1) and instructs the user/agent to run it using the -ExecutionPolicy Bypass flag. This flag is used to circumvent local script execution policies, which is a potential security concern despite the script's intended purpose of verifying repository structure.
[EXTERNAL_DOWNLOADS]: The skill facilitates the installation of external dependencies from public registries. This includes official Remotion framework packages via npx remotion add and the faster-whisper Python library for local audio transcription. These are well-known packages within the video production and AI ecosystems.
[PROMPT_INJECTION]: The skill is designed to process untrusted external data, such as landing page copy, product descriptions, and Figma concepts, to generate video scripts and React components. This ingestion of external content creates an attack surface for indirect prompt injection, where malicious instructions hidden in the source data could manipulate the agent's output.
Ingestion points: SKILL.md and references/production_workflow.md (processing of landing page copy, screen recordings, and product descriptions).
Boundary markers: None identified in the instructions for separating user data from system prompts.
Capability inventory: npx remotion render (execution of generated Node.js/React code), powershell (execution of local validation scripts).
Sanitization: No explicit sanitization or validation of the ingested natural language content is mentioned.

remotion-marketing-video