Client Workflows

Crisis Communication Coach

Drafts a fast, honest crisis communication response for outages, mistakes, or bad news, balancing speed against getting the facts right.

Last updated Jul 11, 2026
FreeClaudeChatGPTCursor
TL;DR

Crisis Communication Coach is a free AI skill for client workflows. Drafts a fast, honest crisis communication response for outages, mistakes, or bad news, balancing speed against getting the facts right. It works with Claude, ChatGPT, Cursor and is ready to use out of the box.

Download Skill.md Package

About this skill

Crisis Communication Coach helps a team respond to a service outage, public mistake, data incident, or other bad-news moment with a message that is fast, honest, and doesn't overcommit to facts not yet confirmed. It separates what's known now from what's still being investigated, avoids corporate-speak that erodes trust, and provides a short follow-up plan so the first message isn't the last communication sent.

What it does

Given a description of the incident (what happened, what's known, what's still unclear, and the audience affected), the skill drafts an initial response message that acknowledges the issue directly, states only confirmed facts, commits to a specific follow-up timeframe, and avoids both panic and defensive minimization. It also outlines what the follow-up message should cover once more is known.

What is included

  • Fast initial-response drafting framework (acknowledge, state facts, commit to follow-up)
  • Known-vs-unconfirmed fact separation to avoid overcommitting
  • Tone calibration to avoid both panic and defensive minimization
  • Follow-up communication plan template
  • Channel-specific variants (status page, email, social post)

How to use it

1. Download the SKILL.md file
2. Upload it to Claude, ChatGPT, or Cursor as a project skill or custom instructions
3. Describe the incident: what happened, what's confirmed, what's still unclear, and who's affected
4. Receive a draft initial response plus a plan for the follow-up communication

Examples

Example input
We had a data breach exposing some customer emails and hashed passwords. We just found out an hour ago and are still investigating scope.
Example output
Initial response draft:
"We're writing to let you know we identified unauthorized access to a system containing customer email addresses and hashed passwords earlier today. We are actively investigating the scope of this incident and have already [contained the access / begun the investigation — confirm actual containment status before sending]. We will share a full update, including specific guidance, within 24 hours. In the meantime, we recommend resetting your password as a precaution."

What NOT to include yet: exact number of affected users, root cause, or any claim that no other data was accessed — these aren't confirmed and stating them prematurely risks a credibility-damaging correction later.

Follow-up plan: within 24 hours, share confirmed scope, root cause (if known), specific remediation steps taken, and any regulatory notifications made, if applicable.

FAQ

What is this skill for?
It drafts a fast, honest crisis communication response for outages, mistakes, or incidents, and plans the necessary follow-up.
How do I install it?
Upload the SKILL.md as a custom skill or instructions file in Claude, ChatGPT, or Cursor, then describe the incident and what's known so far.
Will it help me avoid legal risk in what I say?
It flags claims that shouldn't be made before they're confirmed, which reduces risk, but for incidents with real legal or regulatory exposure (like a data breach), also involve legal counsel before sending.
What if we don't know the full scope of the incident yet?
That's expected — the skill is built to separate confirmed facts from unconfirmed ones and commit to a follow-up timeframe rather than guessing at scope.
How is this different from writing the message myself?
It applies a structure specifically designed to avoid two common crisis-comms failures: overcommitting to unconfirmed facts, and sounding defensive or evasive.
Can it draft different versions for different channels?
Yes — ask for a status page update, customer email, or social post version and it will adapt tone and length accordingly.

Related Skills

Client WorkflowsFree

Client Communication Cadence Manager

Design client communication cadences with updates, meetings, decisions, risks, responsibilities, response times, escalation, and documentation.

ClaudeChatGPT
#client communication#project updates#stakeholder management
Client WorkflowsFree

Client Escalation and Service Recovery Manager

Manage client escalations with severity, ownership, evidence, communication, recovery options, commitments, approvals, closure, and learning.

ClaudeChatGPT
#client escalation#service recovery#account management
Client WorkflowsFree

Client Health and Renewal Review Manager

Create client-health and renewal reviews with outcomes, usage, satisfaction, risks, commitments, expansion, ownership, and follow-up.

ClaudeChatGPT
#client health review#renewal management#customer success

Related Prompts

Free

Automation Failure Monitoring & Recovery

Create an observability and incident-recovery design for business automations, including logs, alerts, retry policy, ownership, replay, and post-incident review.

ClaudeChatGPT
#automation monitoring#incident response#workflow reliability
Free

Customer Review Response & Reputation Manager

Drafts thoughtful responses to customer reviews (positive and negative) that protect brand reputation and demonstrate care to future readers, not just the original reviewer.

ClaudeChatGPT
#review-management#reputation-management#customer-service
Free

Rapid MVP Architecture Decision Framework

Decide on the right technical architecture for a fast-moving MVP build, balancing speed of shipping against future scaling and rework cost.

ClaudeChatGPT
#mvp development#technical architecture#startup engineering