fix: address review feedback from PR #9334#9393
Conversation
Co-Authored-By: Claude <noreply@anthropic.com>
![devin-ai-integration[bot]](https://avatars.githubusercontent.com/in/811515?s=60&v=4){kind=small}
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: f182f8e220
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".
| [GUARDIAN_VERIFY_TEMPLATE_KEYS.TELEGRAM_CHALLENGE_REQUEST]: (_vars) => { | ||
| return 'Vellum assistant guardian verification requested. Reply with the 6-digit code you were given.'; | ||
| [GUARDIAN_VERIFY_TEMPLATE_KEYS.TELEGRAM_CHALLENGE_REQUEST]: (vars) => { | ||
| return `Vellum assistant guardian verification requested. Your code is: ${vars.code}. Reply with this code or use /guardian_verify ${vars.code}`; |
There was a problem hiding this comment.
This change now includes vars.code in every Telegram challenge text, but that template is also used by the numeric chat-ID start_outbound path (assistant/src/daemon/handlers/config-channels.ts), where operators can mistype or reuse a stale destination. Before, a misrouted recipient still could not verify without the out-of-band code shown to the initiator; now the recipient has everything needed in-channel (/guardian_verify <code>) and can bind themselves as guardian, turning destination mistakes into unauthorized binding risk.
Useful? React with 👍 / 👎.
1 participant
Address review comments from #9334 (revert #9333 + rework guardian: show code in UI, reply in channel)