The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use vendor-specific command-line tools, runpodctl and runpod-mcp, to perform infrastructure management tasks such as listing, searching, and deploying pod templates. It also includes example ssh commands for remote file management on the resulting pods.- [EXTERNAL_DOWNLOADS]: The documentation provides an example for downloading a model checkpoint from Hugging Face (huggingface.co) using curl. Hugging Face is a well-known and widely used service for hosting machine learning assets.- [INDIRECT_PROMPT_INJECTION]: The skill utilizes an attack surface where instructions depend on data ingested from external CLI tool outputs.
Ingestion points: Data is retrieved from the output of runpodctl template commands as described in SKILL.md and reference/comfyui.md.
Boundary markers: The skill does not define specific boundary markers or delimiters for the agent to use when parsing the tool output.
Capability inventory: The agent is empowered to deploy infrastructure via runpodctl and interact with remote file systems via ssh.
Sanitization: The skill relies on the agent's internal logic to parse CLI output without providing explicit sanitization or validation rules.

runpod-templates