Find your way.

Operational reference

This guide describes staff responsibilities and safe operating practices. Permissions remain role-based; when a case exceeds your authority, contain immediate harm, document the facts, and escalate.

1. Hierarchy & Responsibilities

The current staff ladder, specialist functions, escalation paths, and role boundaries.

Moderation ladder

@Trainee — learns procedures under supervision; handles low-risk tasks within assigned access.
@Moderator — routine moderation, reports, tickets, evidence collection, and member support.
@Senior Moderator — complex cases, quality review, mentoring, and escalation support.
@Head Moderator — moderation leadership, high-risk decisions, consistency, and team performance.

Administration and management

@Administrator — server administration, operational decisions, configuration, and cross-team coordination.
@Security Administrator — security administration, incident response, access risks, and sensitive investigations.
@Executive Administrator — executive administration and senior operational oversight.
@Executive Manager — executive management, staffing direction, policy, and major escalations.

Specialist and umbrella roles

@All Staff — umbrella role for the staff team.
@Emergency Staff — emergency response for urgent, high-impact incidents.
@Developer — development, technical maintenance, bot behaviour, and implementation support.
@Administrator Permission — privileged-access administration and sensitive access governance.
@Owner — ownership layer and final organisational authority.

Escalate by risk, not ego

Escalate when the case involves credible danger, child safety, doxxing, malware, account compromise, severe hate, staff misconduct, major raids, legal requests, privileged access, or a conflict of interest. Higher roles are not a substitute for documenting what happened.

---

2. Conduct & Staff Standards

Professional conduct, neutrality, confidentiality, boundaries, and use of authority.

Core expectations

Be calm, accurate, proportionate, and respectful. Enforce written policy rather than personal preference. Do not threaten outcomes, argue for entertainment, compete over punishments, or use staff access to win personal disputes.

Conflicts and recusal

Disclose significant friendships, hostility, financial relationships, prior disputes, or personal involvement. Preserve immediate safety if necessary, then hand the substantive decision to an uninvolved qualified reviewer.

Confidentiality

Ticket content, reports, staff discussion, personal data, detection methods, and moderation records are need-to-know. Do not copy them into personal servers, AI services, public posts, or unrelated DMs. Share the minimum necessary for the task.

Member contact

Prefer official tickets and logged channels. Do not request passwords, tokens, authentication codes, recovery codes, or unnecessary identity documents. Do not flirt with, exploit, intimidate, or pressure members through a staff relationship.

Accountability

Record meaningful actions and correct material mistakes. Staff status never exempts anyone from the rulebook. Retaliation against reporters, witnesses, appellants, or colleagues who raise concerns is serious misconduct.

---

3. Moderation Workflow

A repeatable process for reports, evidence, decisions, records, communication, and appeals.

1. Receive and stabilise

Acknowledge the report, identify immediate danger, stop ongoing harm, and protect evidence. Use the least disruptive temporary control that safely contains the situation.

2. Establish facts

Collect message links, IDs, timestamps, original attachments, relevant context, prior connected incidents, and statements. Separate verified facts from inference. Do not conduct public interrogations or ask reporters to obtain dangerous material.

3. Apply policy

Identify the exact rule, assess severity, impact, intent, history, cooperation, evasion, vulnerability, and recurrence risk. Compare analogous cases where available without exposing private records.

4. Act and document

Use the authorised command with the correct target, duration, and reason. Record evidence references, rule basis, responsible staff member, action, and any review or expiry requirement.

5. Communicate and review

Give the affected member the information they are entitled to receive, including duration and appeal route where applicable. Notify relevant staff, monitor recurrence, and correct the record if new evidence changes the finding.

---

4. Lyxia Staff Commands

How staff should use Lyxia moderation, case, role, channel, safety, and administration commands.

Before running a command

Confirm the target, channel, duration, reason, and permission scope. Use Discord autocomplete instead of typing IDs from memory. For destructive or broad actions, pause and verify the request twice.

Common moderation flow

/warn user:@member reason:Reason here — create a formal warning.
/timeout user:@member duration:10m reason:Reason here — temporarily restrict a member.
/case case:123 — inspect a case.
/warns user:@member — review warning history.
/report user:@member reason:Reason here — submit a private report.

Roles and channels

Use /modtools roleadd, /modtools roleremove, and /modtools temprole only when the role is authorised and below the bot/staff hierarchy. Use /lock, /unlock, /slowmode, /modtools lockall, and /modtools unlockall proportionately and document broad changes.

High-impact commands

Bans, mass bans, emergency locking, restore operations, anti-raid, anti-nuke, whitelist changes, and configuration commands require elevated care. Preview or validate when supported; never experiment with them in production.

Complete reference

The Bot Command Reference documents every Lyxia option, example, alias, and permission. Use the Bot Commands button on the Staff Info hub, then open Moderation or Administration.

---

5. Tixia Modmail Commands

Thread replies, saved responses, subscriptions, user controls, snoozing, logs, and Modmail administration.

Replying in a Modmail thread

/reply msg:message text — normal staff reply.
/areply msg:message text — anonymous reply.
/preply msg:message text and /pareply msg:message text — plain variants.
Formatted reply variants support configured variables; verify the preview and recipient before sending.

Thread lifecycle

/close — close now or schedule a close.
/move — move a thread to an authorised category.
/snooze and /unsnooze — temporarily defer and restore a thread.
/repair — repair a thread broken by Discord state.
/loglink and /msglink — retrieve relevant links.

Participants and notifications

Use /adduser, /removeuser, anonymous variants, /notify, /unnotify, /subscribe, and /unsubscribe deliberately. Adding participants may expose thread content; confirm the person belongs in the case before changing access.

Saved replies

The /snippet family creates, edits, removes, previews, selects, and installs reusable responses. Review saved replies regularly so they remain accurate, empathetic, and consistent with current policy. A snippet accelerates writing; it does not replace reading the member’s message.

Configuration

The /config, /permissions, /alias, /autotrigger, /threadmenu, /plugins, and setup families change Modmail behaviour. These are administrative tools, not routine thread commands. Use the Tixia categories in the Bot Command Reference for exact syntax.

---

6. Security & Emergency Response

Urgent containment, evidence safety, compromised accounts, raids, doxxing, and external escalation.

Identify urgency

Treat credible threats, child exploitation, doxxing, swatting, malware, phishing, large raids, active account compromise, and mass permission abuse as urgent. Notify authorised emergency or security leadership without turning the incident into a public spectacle.

Contain safely

Restrict the smallest necessary scope first when practical: remove malicious content, isolate an account, pause invites, lock affected channels, or activate approved protections. Broad actions require continuous review and a clear path to recovery.

Preserve evidence

Save message, user, channel, case, and audit-log identifiers; record timestamps and relevant context. Do not download illegal content, execute suspicious files, click untrusted login links, or spread doxxing material into staff chat.

Compromised accounts

Separate account compromise from malicious ownership where evidence supports it, but protect the community first. Revoke exposed access, advise credential rotation and MFA, review linked actions, and document recovery conditions.

Escalation

Discord Trust & Safety, safeguarding organisations, hosting providers, or lawful authorities may need limited evidence. Only authorised staff should make external disclosures, and they must share the minimum necessary information.

---

7. Communication, Handover & Resources

Internal channels, staff handovers, task ownership, announcements, and operational hygiene.

Staff workspaces

#staff-chat — day-to-day staff discussion.
#administrators — administration discussion.
#staff-commands — staff commands where appropriate.
#staff-info — staff information and reference material.
#staff-announcements — staff announcements.
#staff-tasks — assigned tasks and tracked work.

Handover quality

State what happened, what is verified, what remains uncertain, actions already taken, evidence locations, risk, deadline, and the next responsible person. Do not hand over a pile of screenshots without a concise summary.

Decisions and announcements

Separate proposals, decisions, and implementation notes. Important operational changes belong in the appropriate announcement or reference channel, not only in transient chat.

Access hygiene

Use unique credentials and MFA, lock your device, remove access promptly when roles change, and never share bot tokens or dashboard sessions. Report accidental exposure immediately rather than attempting to hide it.

Ask early

Uncertainty is not failure. Ask a senior staff member before taking a high-impact action when time allows. In urgent situations, contain immediate harm, document the decision, and request review as soon as possible.