| name | region-selector |
| description | Choose cloud regions with cleaner grids — compare deployment regions on carbon intensity alongside latency, data-residency, price, and service availability, for new deployments or migration candidates. Use this skill whenever the user asks which region/zone to deploy in, mentions moving regions, sets up new infrastructure, or asks about the carbon footprint difference between regions. Part of Lean Agentic AI Skills; emits lean-findings.json. |
Region Selector
Advisor skill — moves I at deployment scope. Input: current/candidate regions, latency requirements (user geography), data-residency constraints, services used. Output: lean-findings.json with region recommendations.
Two regions of the same provider can differ several-fold in grid carbon intensity. For workloads without hard locality constraints, region choice is one decision with permanent effect.
Subject type: emit subject.type: "architecture" in findings.
Procedure
- Gather constraints first — residency (legal, non-negotiable), latency budget to user base, required services per region, egress-cost coupling to other systems. Constraints eliminate; carbon ranks the survivors.
- Rank surviving regions by grid intensity. Sources to cite: provider sustainability pages (GCP publishes CFE% per region; AWS and Azure publish renewable/sustainability data), Electricity Maps for the underlying grids, Cloud Carbon Footprint's coefficients. Use current data via web lookup when possible — grids change yearly; date-stamp whatever figures you quote and name the source.
- Distinguish claims honestly: grid carbon intensity (physical, local) vs provider renewable matching (contractual, often annualized) are different things. Prefer hourly/local measures (like CFE%) when ranking; note when a "100% renewable" claim is annual matching.
- Recommend: primary region + rationale; for existing deployments, migration findings only when the win is material and constraints allow (migration effort is real — effort: high, usually).
- Flexible sub-workloads (batch, DR, analytics) can live in cleaner regions even when the primary can't move — often the practical win. Emit as separate findings.
Trade-offs to always name
Latency deltas (state the geography math), egress fees between split regions, service-availability gaps, team familiarity/ops risk, and price differences.
Not this skill's job
Hour-by-hour shifting within a region (carbon-aware-scheduler), sizing (right-size-compute).
