The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses sudo to configure the NVIDIA Container Toolkit (nvidia-ctk runtime configure). This is a standard administrative requirement for setting up GPU-accelerated Docker environments and is consistent with the skill's purpose as an NVIDIA troubleshooting tool.
[EXTERNAL_DOWNLOADS]: Fetches and executes the uv installer from astral.sh (curl -LsSf https://astral.sh/uv/install.sh | sh). uv is a well-known Python package manager, and this installation method is the official recommended procedure.
[COMMAND_EXECUTION]: Performs various system and environment diagnostic checks, including nvidia-smi for GPU status, uname for OS information, and uv pip list for dependency verification. These actions are necessary to fulfill the skill's stated troubleshooting role.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted data in the form of user-provided error messages and full tracebacks.
Ingestion points: User-provided error messages in Step 1 and full tracebacks collected in Step 3 of SKILL.md.
Boundary markers: None identified; user error logs are processed directly to match signatures.
Capability inventory: The skill has access to shell command execution (uname, nvidia-smi), package management (uv sync), and elevated privileges (sudo).
Sanitization: None; the skill assumes the content of error logs is purely diagnostic text.

cosmos3-env-troubleshoot