fix(weixin): preserve voice transcript origin

[zgzczzw]

What does this PR do?

Preserves the voice origin when Weixin supplies a server-side transcript in voice_item.text.

Previously, _extract_text() returned the transcript as plain text. Because _download_voice() intentionally skips downloading audio when Weixin already supplied text, the resulting event had no media and was classified as ordinary text. Downstream agents therefore could not distinguish a voice transcript from text typed by the user.

This change adds a compact platform-origin marker while leaving ordinary text messages unchanged. It is a deliberately narrower alternative to #27235: it preserves the missing semantic information without adding outbound echo state or a new behavior/configuration knob.

Related Issue

Related to #11686 and #27235.

Type of Change

• 🐛 Bug fix (non-breaking change that fixes an issue)
• ✨ New feature (non-breaking change that adds functionality)
• 🔒 Security fix
• 📝 Documentation update
• ✅ Tests (adding or improving test coverage)
• ♻️ Refactor (no behavior change)
• 🎯 New skill (bundled or hub)

Changes Made

• Update gateway/platforms/weixin.py to retain an explicit voice-transcript marker for voice_item.text.
• Add a regression test in tests/gateway/test_weixin.py covering the exact inbound payload shape.

How to Test

1. Send a Weixin voice message for which Weixin provides voice_item.text.
2. Confirm the agent receives [Voice transcription provided by Weixin] followed by the transcript.
3. Confirm typed text continues to pass through unchanged.

Automated verification:

scripts/run_tests.sh tests/gateway/test_weixin.py -v
72 passed, 0 failed

venv/bin/ruff check gateway/platforms/weixin.py tests/gateway/test_weixin.py
All checks passed!

Checklist

Code

• I've read the Contributing Guide
• My commit messages follow Conventional Commits
• I searched existing PRs and issues
• My PR contains only related changes
• I've run the entire pytest tests/ -q suite locally (targeted Weixin suite passes)
• I've added tests for my changes
• I've tested on Linux through a live Weixin gateway

Documentation & Housekeeping

• Documentation update — N/A
• cli-config.yaml.example update — N/A; no config keys changed
• CONTRIBUTING.md / AGENTS.md update — N/A; no architecture or workflow change
• Cross-platform impact considered; change is platform payload text handling only
• Tool descriptions/schemas update — N/A

[tonydwb]

Code Review Summary

Verdict: Approved

Looks Good

• Fix(weixin): preserve voice transcript origin
• 39 additions, 1 deletion — targeted fix
• No issues detected

---

Reviewed by Hermes Agent

[teknium1]

Merged into main via consolidated salvage PR #73515 (merge c911a5f10f). Your Weixin transcript-origin tagging ([Voice transcription provided by Weixin] when no raw audio exists) was cherry-picked with your authorship, merged with #47125.

Your contribution is credited to you in git history. Thank you! Closing this PR as merged-via-salvage.