Language:
This repository is an AI-friendly scaffold project. Agent work in this repository should optimize for explicit structure, low ambiguity, and reusable conventions.
- Prefer shared project conventions over tool-specific shortcuts.
- Treat the CLI as the primary product surface.
- Treat templates as core assets owned by the repository.
- Keep Codex and Claude Code support aligned unless divergence is necessary.
- Prefer readable repository structure over clever automation.
- Treat the repository as AI-authored by default, not manually coded.
This project exists to help users:
- start new projects with agent-friendly structure
- migrate existing projects into an agent-friendly structure
- generate agent collaboration files
- inspect repositories for missing structure and conventions
- Put product intent in
README.md. - Put deeper product decisions in
docs/. - Keep agent-specific guidance minimal and explicit.
- Avoid hidden conventions that are not written in the repository.
- Host integration rules belong in
docs/host-integration.md. - Read current truth before version snapshots:
docs/current-capabilities.mddocs/features/README.mddocs/README.md
- Shared workflow truth lives in
docs/*; skills are thin entry layers.
- Main language: Go
- Primary interface: CLI
- Shared scaffold assets:
templates/ - Implementation code:
cmd/andinternal/
The first stage should focus on repository definition, structure, and documentation before building out a full CLI implementation.
This project is expected to be implemented through AI agents.
- Humans define direction, standards, and acceptance criteria.
- Agents generate and modify repository contents.
- Manual source-code authoring is not the intended operating model.