crisis-respond
Skill17 starsChanged 7 months agoMITView on GitHub
Detect negative comments, trolls, and PR issues early — draft calm, professional responses with clear escalation guidelines for what the agent handles vs. what goes to the founder.
- Reads credentials
In this repository40 other files it gives its agents.
- ab-test-analyzeSkill
- metrics-trackSkill
- performance-reportSkill
- caption-draftSkill
- cta-craftSkill
- hook-writeSkill
- post-createSkill
- thread-createSkill
- apology-draftSkill
- impersonator-monitorSkill
- comment-replySkill
- dm-respondSkill
- engagement-pollSkill
- influencer-outreachSkill
- mention-monitorSkill
- cadence-timingSkill
- content-calendarSkill
- presence-refreshSkill
- profile-auditSkill
- cross-post-adaptSkill
- draft-reviewSkill
- schedule-queueSkill
- thread-structureSkill
- blog-to-threadSkill
- newsletter-to-postsSkill
- podcast-to-clipsSkill
- competitor-content-auditSkill
- prospect-social-auditSkill
- social-profile-discoverSkill
- brand-voice-systemSkill
- content-pillarsSkill
- crm-social-bridgeSkill
- positioning-statementSkill
- news-responseSkill
- trending-briefingSkill
- trend-monitorSkill
- asset-formatSkill
- image-promptSkill
- meme-createSkill
- visual-planSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 63
For trolls: Do not reply. Hide/block. Report if violating policies.
- Distinctive: line 115
Read drafted response. Does it sound like a human who cares?
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Crisis Respond
Detect and respond to negative social situations before they become crises. Speed and tone are everything — a well-handled complaint can build trust; an ignored or defensive response can cause reputational damage that takes months to undo.
Purpose
One hour is the difference between a contained complaint and a viral crisis. This skill triages situations, drafts appropriate responses, and knows exactly when to escalate to the founder.
When to Use
- Negative comment detected on brand post
- Spike in negative mentions
- Customer complaint goes public
- Founder or brand is being criticized
- Potential PR issue brewing
Inputs Required
- The triggering comment, mention, or DM
- Context (one-off or pattern)
- Brand voice guide (for tone calibration)
- Escalation contacts
Quick Reference
| Situation | Response | Escalate When |
| Legitimate complaint | Acknowledge, empathize, move to private | High-value customer, technical issue, refund |
| Fair criticism of content | Thank them, engage respectfully, no defensiveness | Criticism from influential voice |
| Misunderstanding / error | Politely clarify with facts | Clarification might make it worse |
| Troll / bad-faith | Do NOT engage. Hide/report if possible | Impersonating brand or spreading false claims |
| Coordinated attack | Flag immediately. No individual responses. | Immediately to founder |
Procedure
- Detect via
mention-monitor,dm-respond, or manual feed check. - Triage:
- Legitimate complaint or bad-faith?
- One-off or pattern?
- Worst-case scenario if it escalates?
- For legitimate complaints: Acknowledge first. Validate feeling. Offer path to resolution (DM/email). Keep first reply public (others see responsiveness).
- For trolls: Do not reply. Hide/block. Report if violating policies.
- For coordinated attacks: Escalate to founder immediately. Do not reply individually.
- Document every interaction: What triggered, what response, outcome, what to do differently.
Output Format
# Crisis Response: [Situation]
## Detection
**Time:** [Timestamp] | **Platform:** [X / LinkedIn / IG / TikTok]
**Source:** [Comment / DM / Mention / Coordinated]
**Trigger:** [What was said]
## Assessment
**Type:** [Complaint / Troll / Criticism / Attack / Misunderstanding]
**Severity:** [Low / Medium / High / Critical]
**Pattern:** [One-off / Recurring]
## Response
**Drafted reply:** [Full text, tone-calibrated]
**OR no response:** [Why]
## Escalation
**To founder:** [Yes / No]
**Founder briefing:** [Who, what, what's been done, what's needed]
## Monitoring
**Watch for:** [Volume, sentiment, media pickup]
**Check frequency:** [Every 30 min / 1 hour / 4 hours]Done Criteria
The skill is complete when:
- Situation is triaged and categorized
- Response is drafted (or "no response" decision made)
- Escalation decision is clear
- Monitoring plan is set
Pitfalls
- Responding defensively (validates the complaint and escalates)
- Responding too slowly (4-hour-old complaint with 50 replies is harder to contain)
- Engaging with trolls (every reply is fuel)
- Deleting legitimate criticism (Streisand effect)
- Using corporate crisis language ("we take this matter seriously")
Verification
Read drafted response. Does it sound like a human who cares? Would the complainer feel heard? Is it short enough to not be screenshot-worthy? Has the founder been informed if needed?
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

