Repository navigation
fix(glossaries): the agent addresses the person formally; conversational register is for interface text only - #267
Merged
Merged
Conversation
…nal register is for interface text only Spanish's glossary said "conversational tú for UI strings and chat messages", and the pipeline treats glossary rules as binding: the agent's safety exemplars (the agent speaking to a person in distress) came back in tú and passed review at MQM 100, against CIRISAgent's own es guidance (usted whenever the agent addresses the user). Italian's rule even gave an agent line as its tú example. - Twelve T/V glossaries (es, it, de, pt, ru, uk, tr, hi, ur, fa, id, am): the conversational rule now covers interface text (labels, buttons, menus), and a bullet makes the formal address binding when the agent speaks to the person (prompts.*, safety and care exemplars, replies). fr, ja, ko, zh, bn, pa, ta, te, mr, my and th already required the formal form. - TRANSLATION_GUIDE §3 gains "Who is speaking decides the register", so the drafter, the judge and the repair rung all carry it and a reviewer flags a slip. No translated value changes here; values already written in an informal form are retranslated through the evaluate lane. glossary.guidance() carries the rule for all twelve; _guide_rules() parses; pytest testing/ 737 passed. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_019ZeGbVvLGSbRHPJQJA1dao
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
|
You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard. |
…IRISAgent 04_formal_register) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_019ZeGbVvLGSbRHPJQJA1dao
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The Spanish glossary's "conversational tú for UI strings and chat messages" rule leaked into the agent's own speech. Safety exemplars came back in tú and passed review at MQM 100, against CIRISAgent's es guidance (usted when the agent addresses the user). Italian even used an agent line as its tu example.
No translated values change here.
glossary.guidance()carries the rule for all 12,_guide_rules()parses, and pytesttesting/passes (737).🤖 Generated with Claude Code
https://claude.ai/code/session_019ZeGbVvLGSbRHPJQJA1dao