# Discord Admin Elite > Build, harden, and scale elite Discord servers with a practical admin playbook: security baseline, role/permission architecture, onboarding, moderation ops, engagement systems, and analytics-driven iteration. Use when designing a new server, auditing an existing one, fixing chaos, or preparing for growth. - Skill: `modbender/discord-admin-elite` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add modbender/discord-admin-elite` - Raw SKILL.md: https://api.skillmd.com/api/skills/modbender/discord-admin-elite/raw - Safety review: pending (external: skill-scanner PASS, skillspector PASS) - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Marketing & Growth - Author: modbender (https://skillmd.com/u/modbender) - Updated: 2026-09-09 - Page: https://skillmd.com/skills/modbender/discord-admin-elite --- # Discord Admin Elite Use this skill to turn a Discord server into a high-signal, well-moderated, growth-ready community. ## When to use Use when the user asks to: - Set up a new Discord server properly - Improve moderation/safety - Build cleaner channel + role structure - Increase engagement/retention - Audit a messy server and create a fix plan ## When NOT to use Don’t use this for: - One-off Discord message sends/reactions only (use `discord` skill) - Non-Discord community platforms - Bot-specific coding/hosting implementation (use coding workflow) ## Outcomes this skill should produce Always produce these 3 deliverables: 1. **Server Audit Scorecard** (0-100 with category scores) 2. **Priority Fix Plan** (Now / Next / Later) 3. **Execution Checklist** (step-by-step settings/actions) --- ## Core Principles (Elite Admin Standard) 1. **Safety first, growth second** - Lock down abuse vectors before growth pushes. 2. **Least privilege always** - Give members/mods only what they need, no more. 3. **Onboarding is product design** - First 5 minutes determines retention. 4. **Human moderation + automation together** - AutoMod/bots handle speed; humans handle judgment. 5. **Measure, then iterate** - Use Insights and behavior data to decide changes. --- ## Research-backed baseline (use as defaults) ### Security / Anti-raid baseline - Enable **Community** features. - Enable **AutoMod** (keyword/spam/mention abuse filters). - Use higher **verification level** for public servers. - Require **2FA for moderation** roles. - Remove `@everyone` ability to mass mention. - Create a private `#mod-logs` and `#incident-room`. ### Role architecture baseline - Keep a clean hierarchy: Owner > Admin > Mod > Helper > Member > New/Unverified. - Avoid permission overlap chaos (few roles, clear job boundaries). - Never hand out `Administrator` unless absolutely necessary. - Keep bot roles scoped to needed permissions only. ### Onboarding baseline - Configure welcome + rules + clear first action (“Start here”, “Introduce yourself”). - Keep visible channels minimal for new users. - Use concise channel names and category grouping. - Make verification instructions obvious and one-path. ### Engagement baseline - Build 3 core activity loops: 1) Daily discussion prompt 2) Weekly event (AMA, demo, challenge) 3) Recognition loop (wins, shoutouts, roles) - Maintain announcement/value channels users check repeatedly. - Use Insights to evaluate activation + retention changes. --- ## The Elite Server Build Framework (E-SHARP) Use this exact sequence. ### E1 — Evaluate (Audit) Score each 0-20: - Security - Permission hygiene - Onboarding clarity - Moderation operations - Engagement loops Output: total /100 + top 5 risks. ### E2 — Secure - Apply anti-raid + AutoMod + verification + 2FA baseline. - Lock risky perms (`@everyone`, unmanaged bot perms, excessive admin). ### E3 — Hierarchy - Rebuild role map around least privilege. - Document each role: purpose, grants, owner. ### E4 — Activate - Redesign onboarding path for first-message success. - Add one clear CTA channel for newcomers. ### E5 — Retain - Launch weekly cadence (events + content beats). - Create community rituals and recognition. ### E6 — Prove - Track metrics for 2 weeks and adjust: - Join → first message conversion - D7 retention - Moderation incidents/week - Message quality in target channels --- ## Output Templates (copy exactly) ## 1) Discord Server Audit Scorecard - Security: X/20 - Permissions: X/20 - Onboarding: X/20 - Moderation Ops: X/20 - Engagement: X/20 - **Total: X/100** Top Risks: 1. ... 2. ... 3. ... ## 2) Priority Fix Plan ### NOW (24 hours) - ... ### NEXT (7 days) - ... ### LATER (30 days) - ... ## 3) Execution Checklist - [ ] Enable Community + safety features - [ ] Configure AutoMod rules - [ ] Set verification level + 2FA for mods - [ ] Refactor roles and permissions - [ ] Simplify onboarding channels - [ ] Create mod logging + incident channels - [ ] Launch weekly engagement cadence - [ ] Review Insights after 14 days --- ## Suggested Channel Skeleton (starter) ### INFO - #start-here - #rules - #announcements - #roles ### COMMUNITY - #introductions - #general - #wins - #resources ### EVENTS - #events - #event-chat ### STAFF (private) - #mod-chat - #mod-logs - #incident-room - #staff-notes --- ## Common failure patterns (call out hard) - Too many channels too early (overwhelm) - Too many roles with random perms - No verification gate on public invites - No incident playbook for raids - No recurring events/rituals - Measuring vanity metrics only (member count) vs behavior metrics --- ## Trusted references used for this skill - Discord Safety: Auto moderation in Discord https://discord.com/safety/auto-moderation-in-discord - Discord Community: Using Insights to Improve Community Growth and Engagement https://discord.com/community/using-insights-to-improve-community-growth-engagement - Discord Support (raid prevention / roles docs surfaced in research): https://support.discord.com/hc/en-us/articles/10989121220631-How-to-Protect-Your-Server-from-Raids-101 https://support.discord.com/hc/en-us/articles/214836687-Discord-Roles-and-Permissions Note: Some Discord support pages are Cloudflare-protected for automated fetchers; validate in browser when needed.