---
name: discord-webhook-caution
description: Running daily_bot_report.sh or notify_discord.py manually posts a real message to the production Discord webhook
metadata:
node_type: memory
type: feedback
originSessionId: 71d40fa2-b151-4c81-9821-f0dfeb7a0f66
---

`ai-agents/monitor/daily_bot_report.sh` and `notify_discord.py` send to a
**real, hardcoded production webhook** (the vvv-bots health channel). There
is no dry-run flag.

**Why:** I ran `daily_bot_report.sh` manually on 2026-05-15 08:37 to test a
fallback-path edit, after the cron had already fired at 08:00. That posted
a duplicate message to the channel. I should have either skipped the live
send or asked the user first — verifying behavior locally is fine, but
side-effects on shared services need explicit approval.

**How to apply:** To verify monitor changes without spamming Discord:
- Run `summarize_logs.py` alone (`python3 ai-agents/monitor/summarize_logs.py`)
— that only fetches+summarizes, no Discord call.
- For end-to-end behavior, pipe the report into a stub: e.g.
`python3 ai-agents/monitor/summarize_logs.py | tee /tmp/report.txt`
and inspect `/tmp/report.txt` instead of running the cron script.
- Only run `daily_bot_report.sh` in full when the user has agreed to
receive the test message.