Export ML Project

Ordering only. Children own convert and site commands.

Human-facing prose

Details: setup-workspace references/human_facing_prose.md. The multi-select uses Executed notebooks and Documentation site. Do not put skill ids, convert flags, or the wrapper CLI in the question.

Sequence

1. python -m skore_skills status. Read policy.notebooks, policy.site, and skills.

2. AskUserQuestion with allow_multiple. Say first, in 2–4 lines, what each box authorizes (toolchain installs, convert re-execution, a built site) and that both answers persist as policy. A file link is an addition, never the context.

   • Executed notebooks — preselected iff policy.notebooks is not false (true or null)
   • Documentation site — preselected iff policy.site is not false (true or null)

3. Persist each box this turn: checked → python -m skore_skills policy set notebooks true (or site); unchecked → false. Do not leave them null. Do not write notebooks or site into JOURNAL; policy is the record.

4. Load each checked child only if status.skills.<id> is true; else one-line skip. Do not invent that skill's steps:

   • notebooks → export-ml-notebook (installs toolchain if this is the first yes, then convert; --html only if the user wants the notebook viewer on the site)
   • site → export-ml-site (installs mkdocs-material if first yes, init if needed, then build; never convert)

5. Load triage-ml-task only if status.skills.triage-ml-task is true; else stop.

Stop conditions

• Do not git end-turn or git commit.
• Do not invent convert or mkdocs steps when the child skill is missing.
• Do not pixi add / uv add from this coordinator.