agentleFS
kryptophonikkryptophonik/SRE-Skills/.claude/skills/sre-incident-communication/SKILL.md

sre-incident-communication

Skill4 starsChanged 25 days agoMITView on GitHub

Use when an incident update or stakeholder message is needed: announcements, status updates, resolution notices, executive summaries with severity cadence. Drafts only until approved.

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 7 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    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:

    Wait for explicit human approval before any prod-changing command, MCP write tool, or external post.

    line 12
  2. Distinctive:

    Rule: Always state the time of the next update, and send one even if there is no news.

    line 33
  3. Distinctive:

    After approval, hand the exact approved text to the human comms lead, who sends it.

    line 40
  4. Distinctive:

    Use it for any severity. For SEV-3 and SEV-4, the internal announcement is usually the only post (see Update cadence).

    line 47
  5. Distinctive:

    Only for the status page and customer channels.

    line 74
  6. Distinctive:

    Never include secrets, internal hostnames, customer PII, or blame in any message.

    line 119
  7. Distinctive:

    Avoid technical jargon in executive summaries.

    line 120

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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

SRE Incident Communication Templates

Safety gates (REQUIRED)

  1. Read-only diagnosis first.
  2. Propose each change with risk (low/med/high) and reversibility.
  3. Wait for explicit human approval before any prod-changing command, MCP write tool, or external post.
  4. Act only as approved, verify, and stop and escalate if anything gets worse.

Full rules: sre-safety-gates. Read-only MCP setup: sre-safety-gates/references/mcp.md. Draft-only rule: This skill produces drafts. Never post to Slack, status pages, email, or any external channel until a human explicitly approves the exact text and destination. Present each draft with: channel, audience, and text.

Communication channels

  • Slack internal: Primary coordination and high-frequency updates.
  • Slack external: Customer-facing or broad internal announcements.
  • Status page: Formal external status. (always requires approval from IC/comms lead)
  • Email: Executive summaries and post-incident follow-ups.

Update cadence

internal / external; adjust to local policy

  • SEV-1: Internal every 15 min; status page every 30 min
  • SEV-2: Internal every 30 min; status page every 60 min
  • SEV-3: Internal every 2-4 h or on status change; external only if customer-visible
  • SEV-4: On status change only; no external post
  • Rule: Always state the time of the next update, and send one even if there is no news.

Workflow

  • Gather facts (service, severity, impact, symptom, start time, IC, channel). Mark unknowns as [TBD]; never guess.
  • Draft using templates below.
  • Present draft with channel + audience; wait for explicit approval.
  • After approval, hand the exact approved text to the human comms lead, who sends it. Post it yourself only if that human explicitly asks you to, with that text and destination.
  • Schedule next update per the Update cadence section.

Templates

Initial announcement (internal)

Use it for any severity. For SEV-3 and SEV-4, the internal announcement is usually the only post (see Update cadence); skip the status page unless customers can see the impact.

🚨 **NEW INCIDENT DECLARED** 🚨
**Service:** [Service Name]
**Severity:** [SEV-1…SEV-4]
**Impact:** [Who is affected and how, e.g., ~20% of checkout requests fail in EU]
**Symptom:** [e.g., High 5xx rates on Checkout]
**Incident Commander (IC):** [@User]
**Bridge/Channel:** [#inc-checkout-YYYY-MM-DD]
**Status:** Investigating.
**Next Update:** [Time] UTC

Status update

🔄 **INCIDENT UPDATE** ([Time] UTC)
**Current Status:** [Investigating / Identified / Mitigating / Monitoring]
**Severity:** [SEV-1…SEV-4, note any change]
**Impact:** [Current impact; say if it has grown or shrunk]
**Actions Taken:** [e.g., Rolled back v1.2.3, scaling up K8s cluster]
**Next Update:** [Time] UTC

External status page (customer-facing)

Only for the status page and customer channels. No names, IC, bridge or channel links, internal hostnames or internal causes.

**[Investigating / Identified / Monitoring / Resolved]: [Customer-facing feature] [degraded / unavailable]**
**Impact:** [What customers see, e.g., Some customers can't complete checkout]
**Started:** [Time] UTC
**Workaround:** [Steps customers can take, or "None at this time"]
**Next update:** By [Time] UTC

Resolution announcement (internal)

✅ **INCIDENT RESOLVED**
**Service:** [Service Name]
**Resolution Time:** [Time] UTC
**Summary:** [Short description of fix]
**Follow-up:** A blameless postmortem will be scheduled within 48 hours of resolution.

Executive summary (internal)

## Executive Summary: Incident [ID]
**Incident:** [Brief Title]
**Business Impact:** [e.g., Estimated $X revenue lost, Y users affected]
**Root Cause:** [High-level technical cause]
**Short-term Fix:** [What we did to stop the bleeding]
**Long-term Prevention:** [What we are doing to ensure it never happens again]

Output

For each draft, present:

  • Channel and audience: internal Slack, external Slack, status page, or email.
  • Severity: the current SEV, and any change since the last update.
  • Impact: who is affected and how, with [TBD] for unknowns.
  • Text: the filled template, using the external template for anything customers see.
  • Next update: the time, per the Update cadence section.
  • Approval: who must approve before it goes out (IC or comms lead). The draft stays unsent until then.

Best practices

  • Be concise and factual; no speculation on root cause in external posts.
  • Never include secrets, internal hostnames, customer PII, or blame in any message.
  • Avoid technical jargon in executive summaries.
  • Stick to the Update cadence heartbeat.
  • Under-promise and over-deliver on resolution times.

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.