Observability & Incident Response

What it does

Designs production telemetry, alerting, runbooks and incident command so problems are caught and handled fast.

When to use it

Use for structured logs, metrics, alerting design, on-call handoffs, or building real operational readiness.

Problems it prevents

Stops incidents being discovered by a customer complaint instead of an alert that actually fires on the real signal.

Typical workflow

1. Define the outcomes worth observing
2. Instrument safely, without leaking data
3. Alert only on what needs action
4. Run post-incident review

Example output

An alert that fires on a real user-facing signal, with a runbook the on-call person can actually follow at 3am.

What's included

1. 01
   The complete prompt

   The full manual as a ready-to-save SKILL.md file, matched exactly to what ships in the T7 library.

2. 02
   Install walkthrough

   Personal (~/.claude/skills/) and project (.claude/skills/) folder instructions, whichever fits how you work.

3. 03
   Use notes

   How Claude invokes it, what a good result looks like, and what to check before you trust the output.

Works well with

• SystemBackground Jobs & Queues

  Designs reliable background jobs and queues: retries, idempotency, ordering, dead-letter handling, recovery.

• SystemDeployment

  Plans, executes and verifies safe production deployment: build settings, environment config, domains, HTTPS, rollback.

• SystemProject Maintenance

  Keeps launched sites secure, current and reliable through dependency, content, performance and uptime checks.