skills/rehearse-japanese-speaking/references/debrief-and-beta.md

# Debrief and beta mode

Render all headings, explanations, questions, and observations in the configured L1. Keep Japanese reusable expressions in Japanese.

## Standard written debrief

**MISSION**  
What was practiced.

**OUTCOME**  
Choose one:
- Completed independently
- Completed after targeted repair
- Completed with substantial support
- Not yet completed

**WHAT IMPROVED**  
One observable difference between first attempt and replay.

**PRIMARY BOTTLENECK**  
One behavior-level growth area.

**REUSABLE LANGUAGE**  
One or two Japanese expressions only.

**RECOVERY STRATEGY**  
One phrase or interaction behavior.

**LOWER-PRIORITY NOTES**  
Zero to two brief observations.

**NEXT USEFUL PRACTICE**  
One task that targets the same capability.

**RETURN NOTE**  
A compact record:
- task;
- primary focus;
- successful repair;
- remaining difficulty;
- recommended next task.

Avoid vague praise such as “much better.” Describe behavior, for example:

> On replay, you stated the central point first and used one example instead of three.

Do not provide exhaustive corrections or fake numeric fluency/proficiency scores.

## Beta mode

Activate only when the user explicitly requests beta mode, is clearly testing the skill, or provides a beta-testing instruction.

Before the first attempt ask:

> How prepared do you currently feel to handle this task in real life, from 1 to 5?

After the debrief ask:

1. How prepared do you feel now, from 1 to 5?
2. What felt unrealistic, confusing, or unhelpful?
3. Was the primary correction actually the thing that made the task difficult?
4. Would you use this before a real interaction?

Optional research fields:

- first attempt completed;
- spoken repair completed;
- full replay completed;
- realism 1–5;
- correction usefulness 1–5;
- whether the learner later attempted the task in real life.

The strongest qualitative signal is:

> I practiced this, used it in the real interaction, and handled it better than I normally would.