This is a solo learning project (Boostcamp Web & Mobile SNU 2026). External contributions aren't expected, but this file documents how the repo works.
main— deployed. No direct pushes.dev— integration branch.feature/*— one feature per branch, PRs targetdev.
This repo follows Conventional Commits. The Husky commit-msg hook runs commitlint and rejects non-conforming messages.
<type>(<scope?>): <description>
[optional body]
[optional footer]
feat · fix · chore · docs · refactor · test · style · ci · perf · build · revert
editor · preview · theme · storage · print · responsive · deps · ci
feat(editor): add CodeMirror markdown highlightingfix(print): hide toolbar in @media printchore(deps): bump react-markdown to 9.1.0docs(workflow): add multi-agent role separation section
Use the body to explain why, not what. Reference the issue or relevant context.
fix(editor): crash on empty input
CodeMirror state init missing when value === "".
Caught by verifier agent on iOS Safari mobile preview.
- Aim for 50–200 LOC per PR.
- One feature per PR — don't mix theme + storage + layout changes.
- Fill out the PR template (
.github/PULL_REQUEST_TEMPLATE.md). - Attach verifier-agent and reviewer-agent reports (see docs/workflow.md §3).
docs/workflow.md is the source of truth for how I collaborate with AI agents: prompt patterns, verifier / reviewer sub-agents, human checkpoints, anti-patterns.