Review ML Choices
Show stored choices. A change loads the skill that already owns
that decision. Do not write .skore or the journal from here.
Human-facing prose
Details: setup-workspace references/human_facing_prose.md.
The board uses data-science labels: where reports go, executed
notebooks, documentation site, git commits, who manages the
environment, data analysis, and each filled framing cell
(prediction goal, deployment, metric, fold count, and the
others below). Do not put skill
ids, G-* names, or the wrapper CLI in the question.
Sequence
Run python -m skore_skills review choices. The JSON is the
board. Do not rebuild which rows are offered. Do not run
frame show or frame clear here.
AskUserQuestion, one pick. Say first, in 2–4 lines, what a
change authorizes — re-entering that setup, not a silent flag
flip — and echo the current values from the JSON. A file link
is an addition, never the context.
Options are Keep these plus one option per changeable
row and one option per framing row. read_only and
not_offered are context, not options. Use each row's
value. For not_offered, say that row's reason. When
framing_reason is set, say it and do not offer framing rows.
"Who manages the environment" is whether this project manages
it (policy.env.managed). An installed skill is not that
value. Do not write a skill id on the board.
Keep these → stop. Do not load a skill.
One changeable row → load that row's skill and stop. The
close names the decision being re-opened, not the skill id.
Do not policy set. Do not run the child's commands from memory.
skore_mode → the user asked to change where reports go.
notebooks_site → notebooks and the documentation site.
git_autocommit → the user asked to change that choice.
env_managed → do not run env add-skore here. If the
recorded destination is hub or mlflow, do not use
--mode local.
data_analysis when action is rerun → do not write a
JOURNAL skipped row. The written analysis stays; the
change is to run it again.
One framing row → load frame-ml-problem naming that
row's id. Do not blank the cell here.
A request to change a read_only row: say that row's
reason and stop. Do not load setup-workspace to rename.
Do not policy set.
Stop conditions
- Do not
policy set any key.
- Do not edit JOURNAL, including a Data understanding
skipped
row.
- Do not load
sync-ml-reports when policy.skore_mode is
unset.
- Do not
env add-skore, git commit, or scaffold src/.
- If the owning skill is not installed, show the value and skip
the change in one line. Do not invent that skill's steps.
1---2name: review-ml-choices3description: >4 Show stored project choices and re-enter the skill that can change one.5 Trigger when the user asks what we decided, what the current settings are,6 or to change a stored project choice.78 SKIP an explicit sync, export, or “a constraint changed” request — those9 load sync-ml-reports, export-ml-project, or frame-ml-problem directly.1011 HOW TO USE: run `review choices`, show that board, then load one owning12 skill. Do not policy set from here.13metadata:14 source: https://skillmd.com/skills/probabl-ai/review-ml-choices15 modelTier: small16---1718# Review ML Choices1920Show stored choices. A change loads the skill that already owns21that decision. Do not write `.skore` or the journal from here.2223## Human-facing prose2425Details: `setup-workspace` `references/human_facing_prose.md`.26The board uses data-science labels: where reports go, executed27notebooks, documentation site, git commits, who manages the28environment, data analysis, and each filled framing cell29(prediction goal, deployment, metric, fold count, and the30others below). Do not put skill31ids, `G-*` names, or the wrapper CLI in the question.3233## Sequence34351. Run `python -m skore_skills review choices`. The JSON is the36 board. Do not rebuild which rows are offered. Do not run37 `frame show` or `frame clear` here.382. **AskUserQuestion**, one pick. Say first, in 2–4 lines, what a39 change authorizes — re-entering that setup, not a silent flag40 flip — and echo the current values from the JSON. A file link41 is an addition, never the context.4243 Options are **Keep these** plus one option per `changeable`44 row and one option per `framing` row. `read_only` and45 `not_offered` are context, not options. Use each row's46 `value`. For `not_offered`, say that row's `reason`. When47 `framing_reason` is set, say it and do not offer framing rows.48 "Who manages the environment" is whether this project manages49 it (`policy.env.managed`). An installed skill is not that50 value. Do not write a skill id on the board.513. **Keep these** → stop. Do not load a skill.524. One `changeable` row → load that row's `skill` and stop. The53 close names the decision being re-opened, not the skill id.54 Do not `policy set`. Do not run the child's commands from memory.5556 - `skore_mode` → the user asked to change where reports go.57 - `notebooks_site` → notebooks and the documentation site.58 - `git_autocommit` → the user asked to change that choice.59 - `env_managed` → do not run `env add-skore` here. If the60 recorded destination is `hub` or `mlflow`, do not use61 `--mode local`.62 - `data_analysis` when `action` is `rerun` → do not write a63 JOURNAL `skipped` row. The written analysis stays; the64 change is to run it again.655. One `framing` row → load `frame-ml-problem` naming that66 row's `id`. Do not blank the cell here.676. A request to change a `read_only` row: say that row's68 `reason` and stop. Do not load `setup-workspace` to rename.69 Do not `policy set`.7071## Stop conditions7273- Do not `policy set` any key.74- Do not edit JOURNAL, including a Data understanding `skipped`75 row.76- Do not load `sync-ml-reports` when `policy.skore_mode` is77 unset.78- Do not `env add-skore`, `git commit`, or scaffold `src/`.79- If the owning skill is not installed, show the value and skip80 the change in one line. Do not invent that skill's steps.