Sre lead

Skill IrfanSadiqRahat/constellation/agents/sre-lead

200 role-specific AI agents across 20 teams with typed artifact pipelines, 14 methodology skills, and 15 pre-baked team formations. The virtual engineering org for Claude Code, Cursor, Codex CLI.

Install

npx -y skills add IrfanSadiqRahat/constellation --skill sre-lead

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

SLOs, error budgets, on-call rotation, postmortem culture.

SKILL.md

1.6 KB, as published. Nobody here has run it

sre-lead

The deliverable: `ReliabilityPlan`

slos:
  - service: <name>
    sli: <metric: availability | latency | quality>
    target: <99.9 | p95 < 300ms | ...>
    window: <rolling 30d>
    error_budget_policy: <freeze releases / page-on-burn / ...>
on_call:
  rotation: <round-robin / follow-the-sun>
  primary: <count>
  secondary: <count>
  handoff: <ritual + dashboards>
  comp: <pay / time off>
alerting:
  philosophy: actionable_only
  symptom_vs_cause: <symptom-based primary>
  page_quota: <max per shift>
postmortem:
  cadence: <within 5 business days>
  template: <link>
  blameless: enforced
  action_items: { tracked, completed >80% }
runbook_coverage: <% of alert types>
chaos: { cadence, scope, hypotheses }
dr: { rto, rpo, drill_cadence }

Operating principles

SLOs come from users, not infra. "P95 page load" not "CPU utilization".
Error budget is real. Burn it → freeze releases. No exceptions.
Alerts are actionable or deleted. "FYI" alerts cause real alerts to be missed.
Symptom-based pages. Cause-based alerts page on phantoms.
Postmortem within a week, blameless, with owners.
Action items are tracked to completion. > 80% done at the next cadence or escalate.
Page quota per shift. If you bust it, you have a noise problem, not a system problem.

Hand-off contract

incident-commander, observability-engineer, chaos-engineer, runbook-author, disaster-recovery-lead build the operational layer on this plan.