The Agent Skills Directory

[DATA_EXPOSURE]: The diagnostic script scripts/doctor.py reads configuration data from sensitive file paths, specifically .env in the current working directory and ~/.config/video-studio/.env to identify the presence of API keys.
[REMOTE_CODE_EXECUTION]: The skill documentation encourages downloading and executing code from external sources. This includes installing a package via a GitHub archive URL (https://github.com/scrollmark/social-skills/archive/refs/heads/master.tar.gz) and installing the uv tool using a remote shell script (https://astral.sh/uv/install.sh | sh).
[COMMAND_EXECUTION]: The skill uses subprocess.run in scripts/doctor.py to execute shell commands for verifying the status and capabilities of local binaries like ffmpeg and the video-studio CLI. It also provides instructions for the agent to execute system-level setup commands.
[PRIVILEGE_ESCALATION]: The skill describes a video-studio setup --yes command that is capable of performing 'system' level changes, such as using a package manager to install tools like ffmpeg or node outside of the project directory.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from multiple external generative services and stock archives, creating a surface for potential injection from untrusted API responses.
Ingestion points: Environment variables loaded by doctor.py, and outputs from various generative and stock API providers.
Boundary markers: The instructions do not define explicit delimiters or warnings for handling external content.
Capability inventory: The skill has the capability to execute shell commands and trigger system-wide software installations.
Sanitization: There is no evidence of sanitization or validation logic for data retrieved from external service providers.

studio-setup