The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes various uv and dlthub commands to initialize the project environment, add dependencies, and install specific AI toolkits.
Evidence includes shell commands in SKILL.md: uv add "dlt[hub]", uv run dlthub init, and uv run dlthub ai toolkit install.
It also performs local environment inspection using uv run dlthub --version and git state checks.
[EXTERNAL_DOWNLOADS]: The skill references and fetches documentation from official dltHub sources and GitHub repositories to ensure up-to-date configuration.
Targets include official domains like dlthub.com and raw.githubusercontent.com/dlt-hub/.
[DATA_EXFILTRATION]: The skill incorporates strong guardrails against accidental data exposure.
It explicitly instructs the agent to verify that .dlt/secrets.toml is added to .gitignore and never to ask for credentials in chat, directing users to local secret management tools instead.

dlt-bootstrap