agentleFS
gokulsvisiongokulsvision/Crewm8-Social-Media-Manager-Skill-Graph/skills/crisis/apology-draft/SKILL.md

apology-draft

Skill17 starsChanged 7 months agoMITView on GitHub

Draft apology and correction posts with tone calibration for the situation — from minor correction to serious mea culpa — with founder review and approval gating.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    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
  1. Distinctive:

    Every brand will eventually need to apologize.

    line 25
  2. Distinctive:

    No conditional or passive language exists

    line 107
  3. Distinctive:

    Not following through on promised action

    line 117

≈ “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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Apology Draft

Draft public apologies and corrections when the brand needs to address a mistake, misstep, or controversy. An apology done right can strengthen trust. An apology done wrong makes everything worse.

Purpose

Every brand will eventually need to apologize. The difference between a trust-building apology and a reputation-damaging one comes down to three things: accountability, specificity, and action. This skill ensures all three are present.

When to Use

  • Brand posted something with a factual error
  • Brand posted something that offended part of the audience
  • Product/service failure affecting customers
  • Founder made a statement needing contextualization

Inputs Required

  • What happened (full context)
  • Who was affected
  • Severity assessment
  • Founder's willingness to engage

Quick Reference

SeverityApology TypeToneKey Elements
Minor (typo, wrong date)Correction postLight, self-aware"We got [X] wrong — here's the correction"
Medium (insensitive framing)Acknowledgment + apologyGenuine, non-defensive1. Acknowledge impact 2. Apologize without excuses 3. Commit to doing better
Serious (product failure, offensive content)Full apologySerious, accountable, action-oriented1. What happened 2. Full apology (no "if" or "but") 3. What we're doing 4. How we'll prevent it
Crisis (legal, ethical, safety)Crisis statement from founderFounder-led, may require legal reviewFounder's voice, transparent, specific actions

Procedure

  1. Assess severity using the Quick Reference table.
  2. For medium+ severity: Gather full context, identify specific harm. Draft involves founder.
  3. Draft the apology:
    • Minor: 1-2 sentences acknowledging error + correct info + self-aware note
    • Medium: Acknowledge specific mistake → acknowledge impact → apologize directly → state what you'll do differently
    • Serious: State exactly what happened → full unconditional apology → concrete actions → timeline
  4. Avoid apology-killing language:
    • "We're sorry if anyone was offended" (conditional = not an apology)
    • "Mistakes were made" (passive voice = no accountability)
    • "This is not who we are" (if you did it, it is)
  5. Founder review required for medium+ severity. Legal review for serious/crisis.
  6. Prepare monitoring plan: replies to apology, follow-up if needed.

Output Format

# Apology Draft: [Situation]

## Situation Summary
**What happened:** [Specific description]
**Severity:** [Minor / Medium / Serious / Crisis]
**Who was affected:** [People, groups, customers]

## Draft Apology
**Platform:** [Primary platform]

[Full apology text, formatted exactly as it should appear]


## Language Audit
- [ ] Uses "I'm sorry" or "We apologize" (not conditional "if")
- [ ] Acknowledges specific impact
- [ ] No defensive/explanatory language
- [ ] Includes concrete action, not just words

## Review Gate
**Founder review required:** [Yes / No]
**Legal review recommended:** [Yes / No]
**Deadline for posting:** [Timeframe]

## Post-Apology Monitoring
- Monitor for [N] hours
- Watch for: [Continued criticism / Acceptance / New info]

Done Criteria

The skill is complete when:

  1. Severity is accurately assessed
  2. Apology includes specific impact acknowledgment
  3. No conditional or passive language exists
  4. Concrete actions follow the apology
  5. Founder and/or legal review gates are set for anything above "minor"

Pitfalls

  • Apologizing for things that don't need it (legitimizes bad-faith criticism)
  • The non-apology apology (sounds like apology but takes no responsibility)
  • Over-apologizing (10 paragraphs for a typo)
  • Over-explaining (more words = more chances to make it worse)
  • Not following through on promised action

Verification

Strip it down to: "I [did X]. It caused [Y]. I'm sorry. I will [do Z]." If that core is missing, rewrite. Would you accept this apology if directed at you?

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.