You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This repository has an unusually high documentation standard: decisions carry their reasoning, including why the alternatives were rejected, and comments record what was measured rather than what was assumed. It is the project's signature — and its single biggest barrier to contribution, because nobody will guess it.
Without stating it, PRs will arrive at ordinary quality, get heavy review, and the contributor will quietly leave. That is the most common way a promising project loses its first ten contributors.
Acceptance
CONTRIBUTING.md at repo root
The documentation standard, stated with a worked example — take an existing comment from guards.py or market_feed.py and show what makes it acceptable: it says why, it names what was measured, it says what it would take to change the decision
Dev setup: uv sync --all-extras --dev, and the gates a PR must pass — ruff check, mypy, pytest -q (currently 2,762 tests)
TDD expectation: tests first, and evidence in the PR that they failed for the right reason before the fix
This repository has an unusually high documentation standard: decisions carry their reasoning, including why the alternatives were rejected, and comments record what was measured rather than what was assumed. It is the project's signature — and its single biggest barrier to contribution, because nobody will guess it.
Without stating it, PRs will arrive at ordinary quality, get heavy review, and the contributor will quietly leave. That is the most common way a promising project loses its first ten contributors.
Acceptance
CONTRIBUTING.mdat repo rootguards.pyormarket_feed.pyand show what makes it acceptable: it says why, it names what was measured, it says what it would take to change the decisionuv sync --all-extras --dev, and the gates a PR must pass —ruff check,mypy,pytest -q(currently 2,762 tests)Note
Be honest that the bar is high and say why — people will meet a stated bar and resent an unstated one.