Skill repos are git submodules. .openhands/setup.sh initializes them at
session start, before skill discovery runs. In any other context (fresh
manual clone, other agents), initialize them yourself:
git submodule update --init --recursive- agents-summoning — summon other AI agents on PRs/issues: who answers, how to leash a doer, why a summon went silent; source: https://github.com/dogi/agents-summoning
- merge-prepping — rewrite PR titles into house style; source: https://github.com/dogi/merge-prepping
- kotlin-importing — sort/clean Kotlin imports; source: https://github.com/dogi/kotlin-importing
Reviewers speak; doers act — an unleashed doer mention (@openhands, @devin,
@copilot) defaults to commits on your branch, so add "comment only" when that
isn't wanted.
- CLAUDE.md — full codebase guide (architecture, build, conventions)
- docs/DOMAIN_MODEL.md — learning domain: roles, courses, teams, surveys, sync
- docs/CODE_STYLE_GUIDE.md — naming, imports, coroutines, Room, Hilt, UI
- docs/TESTING.md — test patterns per layer