Community Engagement Reply Assistant skill for Claude and Codex
Turns a pile of comments, DMs, and mentions into on-brand replies that sound like a person, with two options for trolls and a flag on anything that needs a…
Install it once and Claude or Codex runs it for you. When your request matches, your AI loads the skill, asks for anything it is missing, and works through the template itself. Prefer to fill it in by hand? See the full prompt template.
Download the skill (.zip) · View SKILL.md
When it is used
Use when a brand's comments, DMs, and mentions have piled up and each needs a reply in the brand voice, including complaints and trolls. Do not use for writing the original posts or captions, or formal support tickets that need account access. Related topics: Marketing, Social Media, Communication.
What it will ask you for
- Brand voice (choose one: Warm and personal, Professional, Playful and witty, Minimal and direct)
- Messages to reply to (longer free text)
- Context (longer free text)
- Message type (choose one: Praise, Question, Complaint, Feature request, Troll or bad faith, Mixed)
- Goal of the replies (choose any: Build loyalty, Move to DM or support, Defuse tension, Encourage sharing, Correct misinformation)
How it works
- Sort the messages by type and platform, so each reply gets the right length and register before any drafting starts.
- Write one reply per message that names the specific thing the person said, confirming none could be pasted under a different comment.
- For each complaint, acknowledge the exact problem first, then give the fix or the route to DM or support with what to send.
- For each troll or bad-faith message, draft a disarming reply and a non-engagement reply, and recommend one with a reason.
- List every message that should be escalated to a human with the reason, so nothing legal, safety-related, or press-related is answered in public.
What you get
One reply per message in the order given, each labelled with the message it answers and the platform, trolls with two options and a recommendation, then a short escalation list with reasons.
What it watches out for
- Replying to a complaint with a generic apology template, which reads as a macro and makes the commenter angrier.
- Arguing facts with a troll in public, which feeds the thread and makes the brand look defensive.
- Promising a feature date in a reply to a request, which becomes a public commitment the product team never made.
- Using LinkedIn-length replies on Instagram, where a paragraph under a comment looks corporate.
How to install
Claude.ai and the Claude desktop app
- Download the skill .zip.
- Open Settings, then Capabilities, and find Skills.
- Choose Upload skill and select the .zip.
- Ask Claude for help with the task. It loads the skill when your request matches.
Claude Code
- Download the skill .zip.
- Unzip it into ~/.claude/skills/ (or .claude/skills/ inside a project).
- Start a new session. Claude Code loads the skill when your request matches.
Codex
- Download the skill .zip.
- Unzip it into ~/.agents/skills/ (or .agents/skills/ inside a project).
- Codex picks it up on its own. Type $ and the skill name to use it, or just ask for the task.
Cursor
- Download the Cursor rule (.mdc).
- Save it in .cursor/rules/ inside your project.
- Cursor's agent applies the rule when your request matches its description.
AGENTS.md (Codex, Cursor, and other coding agents)
- Copy the AGENTS.md section.
- Paste it into AGENTS.md at the root of your repository, creating the file if needed.
- Coding agents that read AGENTS.md follow it when your request fits.