Repository navigation
Home
AgentSustain provides evidence-driven sustainability skills and Python helpers for AI agents. It connects business data, environmental measurements, improvement opportunities, financial models and proposed sustainability actions while preserving evidence and review requirements.
The repository is public and its original project material is MIT licensed. AgentSustain is a development-stage library; public availability does not establish full v1 readiness or professional certification.
- Getting started: installation and runnable fictional examples.
- Preview scope: available capabilities and excluded source packs.
- Evidence and review: provenance, uncertainty and human review.
- Validation: recorded checks and their limits.
- Release decisions: publication history and remaining gates.
For the complete development sequence, read the roadmap. The README covers the seven specialist entry points and links to their contracts. GHG Protocol and Canadian jurisdiction packs are unavailable in this edition; fictional demonstrations do not substitute for real standards or laws.
Development-stage decision-support library. MIT license. Source and professional review requirements remain applicable.