Thanks for contributing to core.
This repository is intentionally small. The goal is not to collect everything related to agent systems. The goal is to preserve a clear, reproducible core.
Good contributions usually improve one of these:
- architecture clarity
- onboarding clarity
- template quality
- default skill quality
- knowledge graph curation rules
- tutorial quality
Please avoid turning this repo into:
- a giant knowledge vault
- a dump of internal runtime logs
- a collection of random agent prompts
- a kitchen sink of domain-specific skills
- a full internal company operating system
Before adding something, ask:
- Does this strengthen the reproducible core?
- Does this encode durable preference rather than temporary workaround?
- Can a new reader understand why this belongs here?
If the answer is no, it probably does not belong in this repo.
- improve the onboarding workflow
- improve the demo agent
- improve templates
- improve the starter knowledge graph structure
- improve the baseline default skills
- improve public-facing documentation
Keep pull requests narrow.
Good PRs usually:
- change one concept at a time
- explain why the change improves the core
- avoid adding unnecessary complexity
- update docs if the architecture or workflow changes
This repo prefers:
- clear boundaries
- small abstractions
- reproducibility
- clean architecture
It does not prefer novelty for its own sake.