llm-patch is community-first. The CLI is the GTM lever; the engine is
a clean library you can extend without forking; and the docs are the
glue.
| I want to… | Channel |
|---|---|
| Ask a question | GitHub Discussions |
| Report a bug | Issues — bug template |
| Propose a feature | Issues — feature template |
| Contribute a plugin (source / generator / registry) | EXTENDING.md + issue template |
| Improve docs | docs issue template |
| Real-time chat | (Discord coming — placeholder) |
llm-patch discovers plugins through Python entry points or the
LLM_PATCH_PLUGIN_<KIND> environment variables (ADR-0008,
EXTENDING.md). The community is encouraged to publish
plugins as separate PyPI packages.
| Plugin | Kind | Author | Status |
|---|---|---|---|
| your plugin here | — | — | open a PR adding a row |
We track work in GitHub Issues with labels:
area:cli— CLI / DX work.area:engine— engine ABCs / pipelines.area:registry— adapter market & hubs.good-first-issue— friendly entry points.help-wanted— actively seeking contributors.
Architectural decisions land as ADRs in docs/adr/. Read those before proposing structural changes; refer to CONTRIBUTING.md for the contribution checklist.
See CODE_OF_CONDUCT.md. TL;DR: be kind, be specific, be patient.