Ready-made coding rules that help your AI agents make the same choices your team would. Hand over routine coding and review to the agents, and keep your team on design and testing.
- Engineering standards for solo developers and small teams, like large companies have
- The same rules for every team member and every project
- Ready-made rules for testing, security, Git, CI/CD and APIs
- Ready-made checks for the rules, run as pre-commit hooks
- Skip, adapt or override the rules
- New rules never change your project without your say
-
Open your project in your agent and paste:
Clone https://github.com/solid-ai-dev/solid-ai-templates next to this project. Read its templates/manifest.yaml, propose the templates that fit this project, and ask me to confirm before going on. Then write AGENTS.md and CLAUDE.md for this project: - Load every file in the manifest's core: list - Load the chosen templates and every file their [DEPENDS ON] lines name - Follow templates/core/agents.md for the output -
Review the result before you commit it.
| Read | For |
|---|---|
| Migrating from 2.x | Moving a 2.x project or fork to 3.0 |
| Specification | How the templates compose: goals, building blocks, rules, quality, risks |
| Adapt the rules | Fork the templates and keep your changes across updates |
| Changelog | What changed in each release |
| Contributing | Setup, and what a change must carry |
| Playbook | Step-by-step procedures |
| Decisions | Why the design is what it is |
| Security | Reporting a vulnerability |
| Issues | Defects, proposals, questions |
MIT. Use, copy and adapt the templates and tools for any purpose, including in closed and commercial projects, keeping the copyright notice. Releases before 3.0.0 remain under CC BY 4.0.
Written by Branimir Georgiev at Imbra. Imbra also offers an extended set of templates, with stacks, platforms and process rules, on request.