The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill utilizes uv run to execute Python scripts and supports piping code directly to the interpreter as seen in references/scripts.md. These are primary, intended functionalities for Python development and are scoped to the user's project context.\n- [EXTERNAL_DOWNLOADS]: The skill manages dependencies via uv add and uv sync, which fetch packages from external registries like PyPI. The instructions emphasize using authorized indexes and existing project configurations, following standard package management practices.\n- [INDIRECT_PROMPT_INJECTION]: The skill has a vulnerability surface for indirect injection as it reads and acts upon external data including pyproject.toml, lockfiles, and script source code.\n
Ingestion points: Project configuration files (pyproject.toml, lockfiles) and Python script files referenced in SKILL.md and references/scripts.md.\n
Boundary markers: None identified in the instruction templates.\n
Capability inventory: Shell command execution via uv run, package installation via uv add, and network operations via uv publish as documented in references/packaging.md.\n
Sanitization: The skill relies on default uv behaviors and does not implement separate validation of project metadata contents.\n- [SAFE]: The skill includes explicit security guidance regarding the handling of sensitive credentials, advising the use of environment variables and CI-based trusted publishing instead of hardcoding tokens in arguments or logs.

manage-python-with-uv