[codex] normalize forwarded Codex commands

[Pigbibi]

Summary

• normalize forwarded Telegram slash commands so /goal.继续... becomes /goal 继续... before reaching Codex
• strip only Telegram bot mentions in the command token, preserving @... text in command arguments
• add regression tests for dotted /goal arguments and bot mention handling

Root cause

The VPS log showed Failed to send keys was emitted after the bot pasted text and pressed Enter, then detected the Codex input row still looked pending. In the reported screenshot the first command used /goal.继续完成验证; Codex expects a space after /goal, so the malformed forwarded slash command could remain pending and surface as a low-level send failure.

Validation

• uv run ruff check src/ tests/
• uv run ruff format --check src/ tests/
• uv run pyright src/telegram_codex_bot/
• uv run pytest --tb=short -q (579 passed)

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 9d27d3ecd4

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 forwarded commands across newlines

When a forwarded command has a multi-line argument, such as /goal.继续\n补充说明 or /goal@CodexBot first\nsecond, this regex does not match because Python's .* stops at the first newline. _normalize_forward_command_text then returns the original text, so the newly added dot normalization and bot-mention stripping are skipped for exactly the kind of long goal/prompt text users can send from Telegram.

Useful? React with 👍 / 👎.