The Agent Skills Directory

REMOTE_CODE_EXECUTION (CRITICAL): The skill contains a pattern that pipes a remote script directly into a shell: curl -LsSf https://astral.sh/uv/install.sh | sh. This allows a remote server to execute arbitrary commands on the system. Per the security policy, piped remote execution from a non-whitelisted source (astral.sh) is a CRITICAL finding.
COMMAND_EXECUTION (HIGH): The skill requires and utilizes elevated privileges via sudo, specifically for restarting system services: sudo systemctl restart myapp. It also modifies system-level files at /etc/systemd/system/.
EXTERNAL_DOWNLOADS (HIGH): The skill initiates network downloads from astral.sh. Since this domain is not within the trusted organization list defined in [TRUST-SCOPE-RULE], it is treated as an untrusted external download.
INDIRECT_PROMPT_INJECTION (HIGH): The skill is designed to ingest and process external project configuration files which can act as a vector for indirect prompt injection.
Ingestion points: pyproject.toml, uv.lock, and other files within the project directory read via Read and Grep tools.
Boundary markers: Absent. There are no delimiters or instructions to the agent to treat file content as untrusted or to ignore embedded instructions.
Capability inventory: The skill possesses extensive capabilities including Bash (shell execution), Write (file modification), and Edit, providing a high-impact exploitation surface if a malicious file is processed.
Sanitization: Absent. The skill provides no mechanisms for sanitizing or validating the content of the project files before they are used in commands or logic.

python-uv