Skip to content

docs(contributing): add structured PR process — issue-first, QA cycle, commit format#85

Open
halindrome wants to merge 3 commits intoDeusData:mainfrom
halindrome:feat/contributing-mr-process
Open

docs(contributing): add structured PR process — issue-first, QA cycle, commit format#85
halindrome wants to merge 3 commits intoDeusData:mainfrom
halindrome:feat/contributing-mr-process

Conversation

@halindrome
Copy link

Closes #84

Summary

This PR documents the contributor workflow first demonstrated in PR #72, which the maintainer described as showing "real engineering discipline." The goal is to give future contributors a clear, reusable playbook for structuring PRs and validating their own work before submission.

Changes to CONTRIBUTING.md:

  • Commit Format section — documents the conventional commit convention (type(scope): description) already used throughout the project history, with a type reference table
  • Pull Request Guidelines — tightened to a focused technical checklist (C code only, one logical change, tests, lint)
  • Pull Request Process section — numbered 4-step submission process: open issue first → describe changes → test against a real project → QA review
  • QA Review Process — Step A/B/C cycle adapted from PR fix(store): keep WAL journal mode during bulk write to prevent DB corruption on crash #72's approach: run a fresh top-tier model session per round, fix findings in a separate commit per round (fix(scope): address QA round N), paste each report as a PR comment to prove the work. Minimum 3 rounds for logic-touching PRs. Docs-only PRs are explicitly exempt.

Notes

This PR is docs-only (changes to CONTRIBUTING.md only) and is therefore exempt from the QA round requirement per the policy being introduced.

@halindrome halindrome force-pushed the feat/contributing-mr-process branch from cc6aac9 to 5bb514c Compare March 20, 2026 21:03
shanemccarron-maker and others added 2 commits March 20, 2026 18:39
…fenced code block

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Stop early when a QA round is already clean — no confirmed critical or
major findings means no further rounds are needed.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

docs: document structured PR/QA review process in CONTRIBUTING.md

2 participants