fix: voice messages silently dropped when STT is not configured

[shaun-agent]

Description

When a user sends a Discord voice message and STT is disabled (stt.enabled = false or unconfigured), the bot silently drops it — no reaction, no reply. The user thinks the bot is broken.

Root cause: Two bugs in src/discord.rs:

1. Audio skip is logged at debug!() level only — operators don't notice it in production logs.
2. Voice-only messages (no text, just audio) pass the empty-check guard because prompt_with_sender always includes the <sender_context> XML block. The agent receives a prompt with only sender context and no actual content, producing confusing or empty responses.

Prior Art

Project	File	Behavior when STT unconfigured
OpenClaw	audio-preflight.ts	Transcribes before mention check; no explicit "STT off" feedback — audio is simply not processed
Hermes Agent	transcription_tools.py	Local/Groq/OpenAI fallback chain; no explicit "STT off" feedback — audio is simply not processed
openab (current)	src/discord.rs	debug!() log, silent skip, empty prompt sent to agent

Neither OpenClaw nor Hermes Agent provides user-facing feedback when STT is unavailable. This is a gap we can fill.

Design

Approach: React with 🎤 emoji + early return for voice-only messages.

Rationale:

• Consistent with existing 🚫 pattern for denied users (react on original message, don't reply)
• Zero channel noise — no extra messages cluttering threads
• Self-explanatory: "I see your mic, but I can't do anything with it"
• Operators get warn!() level logs for visibility

Behavior matrix:

Message type	STT enabled	STT disabled
Voice only	Transcribe → process	React 🎤 → early return
Voice + text	Transcribe → process all	React 🎤 → process text only
Voice + image	Transcribe → process all	React 🎤 → process image only
Text only	Process	Process

Files Changed

File	Change
src/discord.rs	Add warn import, track audio_skipped flag, react 🎤, early return for voice-only

Related

• feat: support voice message STT (Speech-to-Text) for Discord #224 — feat: support voice message STT (design issue)
• PR feat: support voice message STT (Speech-to-Text) for Discord #225 — feat: support voice message STT (implementation, merged)

[CHC-Agent]

Issue at a Glance

User sends voice message (STT disabled)
        │
        ▼
  discord.rs: audio detected, no STT config
        │
        ▼
  debug!() log only → silent skip → user sees nothing 😕

Hi @shaun-agent,

Good UX improvement — the 🎤 react pattern is consistent with the existing 🚫 for denied users. PR #252 is already in progress.