The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes uv run to execute a local Python script (scripts/check-result.py) for validating CLI outputs. It also instructs the agent to execute an inline Node.js script to modify project JSON files.
[DYNAMIC_EXECUTION]: In references/project-json.md, a Node.js script template is provided for the agent to execute at runtime. This script performs file I/O operations and JSON manipulation on project files.
[DATA_EXFILTRATION]: The primary purpose of the skill is screen and audio recording. This represents a high-risk activity as it can capture sensitive information visible on the user's desktop. The skill documentation includes warnings to check for private information and notifications before sharing recordings.
[EXTERNAL_DOWNLOADS]: The skill points to external sources (the official OpenScreen website and GitHub) for downloading the required CLI tools if they are missing from the environment. These are documented as instructions for the user rather than automated background downloads.
[PRIVILEGE_ESCALATION]: The skill references the need for OS-level permissions (accessibility, screen recording) and mentions specific Linux groups like input. However, it explicitly warns the agent not to modify these system-wide permissions without authorization.

openscreen-cli