Make Mirror a continuous language teacher

[pirajoke]

Summary

• continue short learner answers inside the active teaching exercise and prevent repeated examples/tasks
• save catalog words mentioned in Mirror as score-neutral practice history
• show recent practised terms in progress and the latest three consented exchanges in the collapsed Mini App profile
• remove raw provider Markdown and keep application-owned structured markers

Verification

• 1212 unittest tests passed (3 skipped)
• focused Mirror/Mini App contract suite passed
• Python compileall passed
• JavaScript syntax checks passed
• git diff --check passed
• secret pattern scan passed

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-09-16T18:04:50.231809Z	2a6a02f	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 2a6a02f458

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Match CJK terms without treating substrings as words

For Chinese or Japanese packs, this substring check records shorter catalog entries that merely occur inside a longer word. For example, an exchange containing 女人 also matches and persists 人, while 对不起 matches 对 and 不; those extra entries then appear in My Words and as recently practised terms despite never being exposed independently. Use CJK-aware boundaries or a longest-match segmentation strategy rather than testing every compact term with in.

Useful? React with 👍 / 👎.