Skip to content
AgentCustomer Service & Support

Social Media Crisis Response Agent

Free agent kit for when something goes wrong in public: severity level, a first-hour checklist, a short holding statement, a channel plan and what not to say.

Updated Oct 30, 2026Marketplace Founder

About Social Media Crisis Response Agent

When a complaint goes viral or an account gets hacked, the first hour decides how bad it gets. Most small teams have no plan. They stay silent too long, post something defensive, or let a cheerful scheduled post go out in the middle of it.

This free agent kit starts by separating what's confirmed from what's claimed. It rates the situation from level 1 (contained) to level 4 (critical) and says who should lead and when to involve a lawyer, insurer or authorities. It writes a first-hour checklist for your case: pause scheduled posts and ads, save evidence, name one decision-maker, brief staff. It drafts a human holding statement of 80 words or fewer, a response plan by channel, the likely questions with true answers, and a case-specific don't-say list. It finishes with the next 24-72 hours and a review template.

Playbooks cover viral service complaints, food and product safety, hacked accounts, insensitive brand posts, staff conduct on video, false claims and review attacks, and data breaches.

Inside: the agent instructions, a severity matrix with playbooks, statement templates, conversation starters, a Claude Code subagent and a worked example of a viral food-poisoning claim.

For small business owners, social media managers and creators.

Limitations: it isn't legal, PR or medical advice. It can't monitor social media, and it doesn't post anything.

Scanned: no issues found

Release v1.0.0 reviewed by AgentLoomX™ · 9 files · 15.6 KB

Signed release

This package can

Nothing special: the scan found no shell commands, network calls, hooks or access to secrets.

What's inside (9 files)
  • social-crisis-response-agent/CHANGELOG.md
  • social-crisis-response-agent/claude-code/agents/social-crisis-response-agent.md
  • social-crisis-response-agent/examples/example-1.md
  • social-crisis-response-agent/instructions.md
  • social-crisis-response-agent/knowledge/severity-matrix.md
  • social-crisis-response-agent/knowledge/statement-templates.md
  • social-crisis-response-agent/LICENSE.txt
  • social-crisis-response-agent/README.md
  • social-crisis-response-agent/starters.md
Read the README social-crisis-response-agent/README.md
# Social Media Crisis Response Agent

A free agent kit for the worst hour of your social media week. Describe what's happening. It gives back:
- a severity level from 1 to 4, with the reasons;
- a first-hour checklist for your situation;
- a holding statement of 80 words or fewer;
- a response plan by channel, with an owner and a time for each step;
- the likely questions with true answers;
- what not to say, and a plan for the next 24-72 hours;
- a review template for afterwards.

It covers viral complaints, food and product safety claims, hacked accounts, insensitive brand posts, staff incidents on video, false claims and data breaches.

Made by AgentLoomX (https://agentloomx.com). Version 1.0.0.

## Who it's for

- Small business owners who run their own social accounts.
- Social media managers who want a plan ready before they need it.
- Creators dealing with a pile-on or a hacked account.

## What's inside

```
social-crisis-response-agent/
  instructions.md                    the agent's instructions (system prompt)
  knowledge/severity-matrix.md       levels 1-4, the first-hour checklist, playbooks by scenario
  knowledge/statement-templates.md   holding, apology, correction, update and closing templates; don't-say list; review template
  starters.md                        conversation starters
  claude-code/agents/social-crisis-response-agent.md   Claude Code subagent version
  examples/example-1.md              worked example: a viral food-poisoning claim against a restaurant
  README.md, CHANGELOG.md, LICENSE.txt
```

## Setup

### ChatGPT (GPT Builder)
1. In ChatGPT, go to Explore GPTs, then Create.
2. Open the Configure tab. Name it "Social Media Crisis Response Agent".
3. Paste the full text of `instructions.md` into Instructions.
4. Under Knowledge, upload the files in `knowledge/`.
5. Add the conversation starters from `starters.md`, then save.

### Claude (Projects)
1. In Claude, create a new Project.
2. Paste the full text of `instructions.md` into the project instructions.
3. Add the files in `knowledge/` to the project knowledge.
4. Start a chat in the project.

### Google Gemini (Gems)
1. In Gemini, open Gems and create a new Gem.
2. Paste the full text of `instructions.md` into the instructions box.
3. Add the files in `knowledge/` as knowledge files, then save.

### Claude Code
1. Copy `claude-code/agents/social-crisis-response-agent.md` to `~/.claude/agents/` (all projects) or `<project>/.claude/agents/` (one project).
2. Optional: copy the `knowledge/` folder into your project, for example `docs/social-crisis-response-agent/knowledge/`. The subagent looks for `**/social-crisis-response-agent/knowledge/*.md`.
3. Ask: "Use the social-crisis-response-agent subagent to plan our response. The facts so far are in incident/notes.md."

## Usage examples

- "A customer's TikTok about us has 200k views. Here's what happened."
- "Our Facebook page was hacked overnight. What do we do first?"
- "Write an apology for the post we put up this morning."
- "Help me fill in the review now that it's over."

See `examples/example-1.md` for a full run.

## How to customize

- **Prepare now:** fill in your contacts (owner, lawyer, insurer, local health department or regulator) at the top of `knowledge/severity-matrix.md`.
- **Your channels:** list your accounts and who has the logins, so the checklist covers every channel.
- **Your voice:** adjust the templates in `knowledge/statement-templates.md` to sound like you, keeping the structure.

## Limitations

- It isn't legal, PR or medical advice. At level 3-4, involve a lawyer and your insurer, and the authorities where required.
- It can't see how far something has spread. It relies on what you tell it.
- It doesn't post, delete or message anyone.

## Support

Questions or problems? Use "Ask a question" on this product's AgentLoomX listing page.

## License

AgentLoomX Standard License. See `LICENSE.txt`.

SHA-256 c619e8d8764cf64bf7f7fbe6ef7d6ea751bf960185837981c6c155fd85e69a2c

Signed with Ed25519 key a9c62a7921065aa9. Public keys

Scans catch known risky patterns; they are not a guarantee. Report anything that looks wrong to support.

Highlights

  • Severity levels 1-4 with who should lead
  • First-hour checklist for your situation
  • Holding statement of 80 words or fewer
  • Channel plan with owners and times
  • Likely questions with true answers
  • Playbooks: hacks, safety claims, bad posts
  • crisis management
  • social media
  • reputation
  • holding statement
  • community management
  • pr