A Claude Code agent and skill that embeds Green Software Foundation sustainability practices directly into your development workflow.
| File | Type | Purpose |
|---|---|---|
.claude/agents/green-software-engineer.md |
Agent | On-call sustainability consultant for Q&A, architecture advice, and deep code audits |
.claude/skills/green-software/SKILL.md |
Skill | Passive green review that auto-triggers when writing backend code |
| Page | Description |
|---|---|
| Agent Guide | How to invoke @green-software-engineer — use cases, invocation syntax, and knowledge base |
| Skill Guide | How /green-review works — auto-trigger rules, checklist categories, and output format |
- Claude Code v1.0 or later — available as a CLI, desktop app, or IDE extension (VS Code, JetBrains)
- A Claude Code subscription (no additional API keys required)
your-project/
└── .claude/
├── agents/
│ └── green-software-engineer.md
└── skills/
└── green-software/
└── SKILL.md
Or for personal use across all projects, place them under your global config:
~/.claude/
├── agents/
│ └── green-software-engineer.md
└── skills/
└── green-software/
└── SKILL.md
claudeAsk the agent a question:
@green-software-engineer what's the greenest way to handle image uploads at scale?
Run an explicit green review:
/green-review
Let the skill work passively — write a new API endpoint or database query and Claude will automatically apply the green review checklist before responding.
A senior green software engineer persona grounded in:
- The 8 GSF Principles — GSF (Green Software Foundation) defines eight pillars: Carbon, Electricity, Carbon Intensity, Embodied Carbon, Energy Proportionality, Networking, Demand Shaping, and Measurement & Optimisation
- The SCI Standard (ISO 21031:2024) — SCI (Software Carbon Intensity) is a formula for measuring how much carbon a software system emits per unit of work:
SCI = ((E × I) + M) / R - Green Software Patterns from patterns.greensoftware.foundation
Example invocations:
@green-software-engineer review this API handler for carbon efficiency
@green-software-engineer how would I calculate an SCI score for this service?
@green-software-engineer is polling better or worse than websockets from a green perspective?
@green-software-engineer audit this Terraform config for over-provisioning
→ Full documentation in docs/agent-guide.md
A structured sustainability review checklist that covers:
- ⚡ Energy Efficiency — loops, lazy vs eager loading, polling patterns
- 🌐 Network Efficiency — payload size, round trips, CDN usage
- 🗄️ Compute & Storage — query optimisation, caching, data retention
- ☁️ Cloud & Infrastructure — provisioning, scale-to-zero, region selection, time-shifting
- 🔋 Hardware & Embodied Carbon — utilisation, managed vs self-managed
Auto-triggers when writing: API endpoints, database queries, cloud config, background jobs, data pipelines, caching logic.
Severity ratings: 🔴 High / 🟡 Medium / 🟢 Low / ✅ Already Green
→ Full documentation in docs/skill-guide.md
| Agent | Skill | |
|---|---|---|
| Invoke with | @green-software-engineer |
Auto, or /green-review |
| Best for | Q&A, architecture, deep audits | Passive guardrail while coding |
| Runs in | Isolated context window | Your main conversation |
| Output | Conversational, advisory | Structured checklist |
