# Back for More

> A playful habit-loop reader for one moreish everyday thing - a crossword, a shop window, a biscuit tin. Describe
> its four stages (trigger, action, reward, investment), rate how hard each pulls from none to strong, and the page
> draws a carousel with one horse per stage riding higher the harder it pulls, adds up the pull out of 12, picks the
> strongest stage and gives a verdict. A metered run adds a ride report. Built on the SkillSafe skill
> @pmprompt/hooked-model.

https://back-for-more.skillsafe.ai/

## What is free
- The loop form (thing, what you want, four stages each with a note and a pull of none/faint/firm/strong, plus cue type for the trigger and reward kind and variability), .json drop, the carousel card, the total pull, the strongest stage (ties go to the earlier stage) and the verdict all run in the browser.
- Verdicts, in check order: just watching (total 3 or less), one more go (any stage at none, or total 4 to 6), round again (7 to 9), all the way round (10 to 12).
- Four sample reports replay at no cost, one per verdict.
- Downloads: the card (SVG, PNG), the stages (CSV), the full reading (Markdown, JSON).

## The metered task
One task, `ride`, on gpt-terra, metered per run. The input is string fields computed by the page: `task`, `thing`, `want`, `loop` (four numbered lines), `total`, `strongest`, `level_with`, `gaps` and `verdict`. The reply is one JSON object: `verdict` (the page's, copied), `headline`, `stages` (four {stage, note} in loop order), `strongest`, `move` and `caveat`. The page re-checks the reply against its own reading.

A light reading of the person's own notes, not professional advice. Anything involving gambling, drinking, smoking or harm is pointed to proper support.

## Links
- API tutorial: https://back-for-more.skillsafe.ai/api.html
- Source skill: https://skillsafe.ai/skill/@pmprompt/hooked-model (pmprompt/claude-plugin-product-management)