Revise contribution guidelines for clarity and updates - #932
Conversation
Updated the contributing guidelines to reflect the current backlog and number of evaluation cases in a more generic way.
The two clusters (tool / adapter bridges, agent-CLI runtime adapters) are the part a newcomer navigates by, and they do not go stale. The counts and the #301-#322 ranges were the stale part -- #301 and #313 are already closed. This drops every number and issue reference so nothing here can rot again, while keeping the technology lists that let someone self-select a first issue. Generated-by: Claude Code (Opus 5)
potiuk
left a comment
potiuk
left a comment
There was a problem hiding this comment.
The eval-fixture correction here was overdue, and removing the hardcoded issue ranges was the right instinct. I verified all four numbers and then pushed one commit to this branch (b74a1e2) to restore something the diff had traded away — details below, so please sanity-check it rather than take my word for it.
What I checked against the tree and the tracker:
| Claim | Actual |
|---|---|
old: 380+ cases across 18 skills |
1102 case-* fixtures across 70 suites, one per skill — badly stale |
new: 1000+ cases across all skills |
accurate |
old: 24+ such issues |
60 open good first issue — stale, and understated |
#301–#312 / #313–#322 ranges |
#301 and #313 are closed, #312 and #322 still open — the ranges no longer map cleanly, so dropping them was correct |
What I changed, and why
The current backlog has several such issues. undersold a 60-issue onboarding surface, and the same deletion removed the two cluster descriptions — which were the most navigable part of that paragraph for a newcomer, because they name concrete technologies someone can recognise and self-select against (Bugzilla, GitLab, Matrix, Cursor, Aider, …).
The stale things were the counts and the issue numbers. The categories are stable. So b74a1e2 keeps the clusters and drops every number and every issue reference, which means nothing in that paragraph can go stale again:
The label link above is always current, and the backlog is
substantial. Two broad clusters recur:
- **Tool / adapter bridges** — JIRA write path, Bugzilla, IMAP / mbox
concrete wiring, GitLab, Mailman 3 / Hyperkitty, Discourse, Zulip,
Matrix, Forgejo, OSV.dev, Pagure.
- **Agent-CLI runtime adapters** — Codex, Gemini, local-LLM, Cursor,
Aider, gh-copilot, Goose, Amazon Q, Junie, OpenHands.The list indentation matches the pattern that was in main before this PR, so prek is green on the new head — markdownlint, lychee, and the doctoc check all pass, and no heading changed so the TOC is untouched.
If you would rather keep that paragraph short, revert b74a1e2 and merge — the eval-count fix alone is worth landing, and I would not block on this.
This review was drafted by an AI-assisted tool and
confirmed by an Apache Magpie maintainer. The maintainer
approving this PR has read the findings and signed off. If
something feels off, please reply on the PR and a maintainer
will follow up.More on how Apache Magpie handles maintainer review:
CONTRIBUTING.md § Opening a pull request.
2 participants
Summary
Updated the contributing guidelines to reflect the current backlog and number of evaluation cases in a more generic way.
Type of change
.claude/skills/<name>/) — eval fixtures updated belowtools/<system>/*.md)tools/*/withpyproject.toml)docs/,README.md,CONTRIBUTING.md)projects/_template/)prek, workflows, validators)Test plan
prek run --all-filespassesuv run pytest/ruff check/mypypasses(
PYTHONPATH=tools/skill-evals/src python3 -m skill_evals.runner tools/skill-evals/evals/<skill>/)(a regression test for the bug fixed / the behaviour added — see CONTRIBUTING.md)