Community Platform CLI

What it does

Defines a workflow to manage a forum/school-style online community from an agent or CLI: read the feed, reply to comments, publish posts with files or polls, build class lessons, and answer member DMs. Actions that publish, send, or delete anything require an explicit confirmation flag and never run by default.

When to use it

• When you manage a paid community and want to review activity without opening the web UI each time.
• When you need to reply to member comments or DMs faster with assisted drafts.
• When publishing recurring educational content (class lessons) and you want a repeatable process.

Steps

1. Store community access credentials as environment variables, never in plain text.
2. Read the feed and recent comments in read-only mode by default.
3. Draft the proposed reply or post.
4. Present the draft for human approval before any write action.
5. Execute publish, reply, or delete only with an explicit confirmation flag.
6. Log every write action with date and outcome for auditing.

The skill

---
name: community-platform-cli
description: Manages an online community (feed, comments, posts, classes, DMs) from an agent or CLI, with explicit flags for every write action.
---

# Community Platform CLI

## What it does
Lets you read and operate a forum/school-style online community from a coding agent, covering feed, comments, posts with files/polls, class lessons, and member direct messages.

## When to use it
Use it to manage paid or free communities (including GoHighLevel client communities) when you want faster replies and posting without losing human control over what gets published.

## Authentication
1. Use the platform's session or unofficial API with your own credentials, stored only in environment variables.

## Reading
2. By default every operation is read-only: reading feed, comments, and DMs.

## Drafting
3. Every reply, post, or lesson is drafted first and shown for review.

## Explicit confirmation
4. Publishing, replying, deleting, or sending a DM requires an explicit flag like --confirm; without it, nothing executes.

## Audit trail
5. Every write action is logged with date, type, and outcome.

## GoHighLevel angle
6. Qualified conversations from the community can be flagged for manual handoff into the GHL pipeline, without automating sales message sends.

## Safety rules
1. Every deliverable (script, email, post, ad, design) is produced as a draft first.
2. Nothing is sent, published, or paid for without explicit human approval.
3. Never invent metrics, testimonials, customer names, or data not present in the provided sources.
4. No credential, API key, or token is ever written or pasted into chat; reference environment variables instead.

For Claude Code, save it as ~/.claude/skills/community-platform-cli/SKILL.md. In Cursor, ChatGPT or another assistant, paste it as project instructions.