The Agent Skills Directory

[SAFE]: The skill serves as a reference and instruction set for the dlab CLI tool, focusing on the creation and management of 'decision-packs' for data science agents. It does not contain malicious code or instructions.
[SAFE]: The documentation explicitly promotes security best practices, such as running agents in frozen Docker environments and managing credentials using .env files and environment variable forwarding.
[SAFE]: The methodology rules (e.g., 'Never Fabricate', 'Understanding over fitting') are designed to ensure agent reliability and integrity in analytical workflows.
[COMMAND_EXECUTION]: The framework described by the skill facilitates the execution of Python and shell code inside Docker containers, which is the intended functional purpose for data science tasks.
[EXTERNAL_DOWNLOADS]: The skill documents connections to official vendor-managed endpoints, specifically hub.decision.ai for skill management and models.dev for model configuration.

dlab-cli