This repository defines the methodology standard for agent-assisted software development. It covers artifact structure, workflow execution, and run scope/permissions.
- Read manual/overview.md for scope, terms, and operating surfaces.
- Read manual/spec/target-layout.md for target repo structure.
- Read manual/spec/entry-document-contract.md for methodology entry-document requirements.
- Read manual/spec/workflow-contract.md for runtime entrypoint requirements.
- Read manual/guides/claims-and-evidence.md for the claims-first model: which trust-relevant run information becomes claim types and how those claims are packaged as evidence during a run.
- Read manual/spec/evidence-contract.md for evidence terminology, layout, and contract requirements for the runtime evidence surface.
- Read manual/guides/designing-a-workflow.md for workflow design principles and patterns.
- Read manual/spec/artifacts-contract.md for artifact inventory requirements.
- Read manual/spec/labels-contract.md for label behavior requirements.
- Read manual/spec/library-contract.md for library requirements.
- Read manual/spec/status-contract.md for source tracking and permission contracts.
- Read manual/spec/setup-sequence.md for setup order.
- Read manual/guides/semantic-validation-process.md for semantic audit procedure.
- Use manual/guides/authoring-a-methodology.md when creating or updating a methodology source repo.
- Use AGENTS.md for repo-level operating instructions when running agents in this repository.
Please see our Contributing Guide for details.
Run scripts/validate.sh before opening a pull request.
Source files should include SPDX-License-Identifier: (CC0-1.0 OR 0BSD) where
applicable.
This project is released under a dual-license model. Choose either:
This applies to all project materials.