The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill references official installation scripts for the uv tool from the vendor's official domain.
Evidence: curl -LsSf https://astral.sh/uv/install.sh | sh and powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex" in SKILL.md.
Note: These installation methods are documented as the official entry points for the Astral uv project.
[PRIVILEGE_ESCALATION]: Documents the use of execution policy bypass for the Windows installation script.
Evidence: -ExecutionPolicy ByPass in the PowerShell installation example (SKILL.md).
Note: This is standard practice for installing developer tools via PowerShell when the local environment has restricted script execution.
[COMMAND_EXECUTION]: Provides instructions for standard dependency management and tool execution within virtual environments.
Evidence: uv run, uv sync, and uv add commands in SKILL.md.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted configuration files and provides the ability to execute environment-specific code.
Ingestion points: pyproject.toml, uv.lock (SKILL.md).
Boundary markers: None.
Capability inventory: uv run, uv add, uv sync (SKILL.md).
Sanitization: The skill relies on the standard validation logic of the uv binary.

uv