Discord Ban Appeal Review Assistant

Gives the mod team a fast, evidence-based starting point for every appeal. The member submits an appeal via a web form; the agent retrieves their complete strike and case history, reads the appeal text, and produces a structured recommendation with the reasoning and the relevant prior incidents attached — so mods vote with full context instead of from memory.

Use it when appeals pile up or get decided inconsistently, or when the deciding mods weren't involved in the original sanction. It keeps appeals fair and fast without auto-deciding anything — humans still cast the final vote.

1. 1A member submits the appeal form, firing a webhook with their ID and statement.
2. 2The agent fetches that member's full case history from the Notion case log.
3. 3An OpenAI step weighs the appeal against the offense severity and prior pattern.
4. 4A logic branch flags weak appeals for fast dismissal versus substantive ones for review.
5. 5The agent posts a recommendation card (uphold / reduce / overturn) with cited history to the mod channel for a vote.