| name | community-ops-router |
| description | Use when diagnosing, designing, coordinating, operationalizing, or reviewing an end-to-end community operations system; when a request spans onboarding, activities, creators, paid members, reports, automation, support risk, operator writing, SOPs, or AI/bot/operator handoffs; or when the correct specialist community-ops skill is unclear. Produces a routed workflow with user segment, friction, next action, surface, evidence, execution status, feedback loop, review rhythm, and approval boundaries. |
Community Ops Router
Use this skill as the entry point for cross-functional community operations work.
Project Inputs
The kit supplies reusable judgment patterns, workflows, output structures, and safety boundaries. It does not supply a community's names, platform setup, channels, roles, activity rules, rewards, member benefits, policies, user records, credentials, or bot configuration.
Collect only the project inputs needed for the current task. Treat repository examples as demonstrations, never as defaults. If a necessary value is missing, continue with confirmed facts where useful and label the value data gap, [owner to confirm], or a question. Ask only when the missing value blocks a safe or valid result. Never request or preserve passwords, tokens, webhook URLs, full private messages, payment details, or unnecessary personal data.
Operating Model
community goal
-> user segment
-> current friction
-> next user action
-> surface
-> proof/evidence
-> feedback, support, recognition, or reward
-> review and next decision
Start with the user path and operating evidence. Do not begin with channel decoration, generic engagement tactics, or a long list of content ideas.
Read operating-principles.md when the request contains messy records, competing priorities, several possible interventions, or a need to explain why an action was chosen. Apply its evidence, minimum-change, surface-role, automation-boundary, and review rules before routing specialist work.
First Move
Classify the request:
- Diagnose: find the main break in the community lifecycle.
- Design: build an operating loop, user path, or coordinated plan.
- Route: send a concrete task to the smallest specialist skill.
- Operationalize: turn a decision into a reusable SOP with ownership, handoffs, status, verification, and writeback.
- Review: combine evidence, risks, decisions, and next actions across skills.
For multi-step work, define a lean task contract: goal, relevant context, hard constraints, required evidence, success criteria, output format, approval boundary, and stop/question conditions. See task-contract.md.
When another operator, AI agent, bot, scheduler, or integration will carry out the action, use sop-contract.md. The task contract defines the assignment; the SOP contract defines repeatable execution and proof.
Skill Routing
- activity, campaign, challenge, submissions, voting, rewards, reminders, or results ->
community-ops-activity;
- onboarding, welcome, routing, first action, activation, or early retention ->
community-ops-onboarding;
- creators, contributors, UGC, showcases, tutorials, cases, attribution, or reusable content ->
community-ops-creator-assets;
- paid, premium, VIP, high-value, benefits, renewal, entitlement, or priority support ->
community-ops-paid-members;
- daily work log or daily recap ->
community-ops-daily-report;
- weekly metrics, retrospective, decisions, or next-week priorities ->
community-ops-weekly-review;
- support, moderation, privacy, payment, access, permissions, rewards disputes, bots, or official claims ->
community-ops-support-risk;
- recurring collection, drafting, reminders, reports, checks, handoffs, or audit trails ->
community-ops-automation-workflow;
- announcements, reports, articles, stakeholder writing, or natural operator voice ->
community-ops-expression.
If several areas are involved, keep this router as the synthesis layer and use specialist skills only for their part of the work.
Workflow
1. Establish The Operating Context
Identify:
- community type and goal;
- target user segment;
- lifecycle stage;
- available surfaces and tools;
- confirmed evidence and data gaps;
- public, financial, permission, safety, or privacy boundaries.
Do not require the user to complete every possible project field before starting. Gather the smallest relevant input set, then expand it only when the routed task needs more detail.
Separate the input into four buckets before drawing conclusions:
- confirmed facts;
- working assumptions;
- data gaps;
- actions waiting for approval.
Use community-patterns.md for transferable hypotheses when project evidence is limited. These patterns may suggest questions or options; they must not fill project-specific fields.
2. Find The Main Friction
Map the lifecycle:
discovery -> entry -> first action -> repeat participation -> contribution -> recognition -> retention/upgrade -> review
Choose the smallest useful intervention. Name where it happens and what proof shows whether it worked.
3. Route And Produce
For each routed task, require:
trigger -> input -> operator judgment -> output -> evidence -> review rule
Keep confirmed actions, drafts, assumptions, and missing data separate.
4. Operationalize The Handoff
For repeated or delegated work, define:
- trigger and decision owner;
- required inputs and durable evidence source;
- operator, AI, automation, and platform-adapter roles;
- current state and allowed state transitions;
- approval scope and stop conditions;
- execution receipt, verification, and writeback destination.
Use these normal states:
observed -> scoped -> prepared -> pending approval -> approved
-> executed -> verified -> recorded -> reviewed
Use blocked, data gap, ambiguous, or cancelled when the normal path stops. Never report prepared, approved, or executed but unverified as completed.
5. Close The Loop
Every plan should specify:
- one next user action;
- feedback, support, recognition, reward, or follow-up;
- evidence collected;
- review timing;
- decision owner;
- what changes after the review.
Approval Boundary
The agent may read, organize, compare, calculate, draft, and prepare checklists when allowed. Require explicit confirmation before public posts, DMs, user mentions, rewards, discounts, paid-benefit promises, role or permission changes, bot or webhook changes, winner lists, deletions, or official claims.
Output Pattern
**Diagnosis**
[The main operating friction and affected segment.]
**Target Loop**
entry -> action -> proof -> feedback/support -> next action -> review
**Routed Actions**
| Priority | Specialist skill | Action | Surface | Evidence | Owner gate |
|---|---|---|---|---|---|
**Data Gaps**
- ...
**Success Criteria**
- ...
**Next Review**
- Timing: ...
- Decision: ...
Quality Bar
Reject plans that rely only on message volume, vague engagement goals, or unsupported growth claims. Prefer concrete actions, visible evidence, clear ownership, and a repeatable review loop.
