Strengthen contributor guidance to support Help-Wanted contributions

Overview

We are increasingly labeling small, well-scoped features as Help-Wanted (for example #1003 and #1058) to invite community contributions. To set new contributors up for success - and to keep the quality bar high - our contributor guidance needs to clearly communicate our expectations before someone starts work.

This is a tracking issue for a set of documentation and process improvements. Each sub-issue can be addressed by its own PR that closes it, moving us toward the broader goal.

Goals

• Give new contributors clear, discoverable orientation to the project and our expectations.
• Set explicit expectations for tests, coding conventions, and evidence that a change works.
• Establish a guardrail so we are not opening the floodgates to low-effort AI-generated PRs, while still welcoming responsible AI-assisted contributions.
• Reinforce the existing "to spec or not to spec" flow for Issue-Feature work.

Work items

• Define testing requirements in CONTRIBUTING.md (sub-issue)
• Add coding conventions guidance for contributors (sub-issue)
• Add a policy on AI-generated / AI-assisted contributions (sub-issue)
• Update the pull request template: evidence, tests, and AI disclosure (sub-issue)

Engineers will be looped in on the sub-issues where the team needs to define or ratify specifics (e.g., the exact testing bar, the coding conventions themselves, and the wording of the AI policy).

Follow-up (dependent on the above)

Once the guidance is in good shape, add a policy bot rule (.github/policies/) so that when the Help-Wanted label is applied to an issue, a concise orientation comment is posted linking contributors to the updated guidance (spec-or-not, tests required, evidence expected, AI policy). Tracked separately once these land.

(Created with GitHub Copilot assistance.)

Start with CONTRIBUTING.md and the pull request template, then review the listed work items and the existing “to spec or not to spec” flow. The issue requires defining or ratifying testing, coding-convention, AI-assistance, evidence, and disclosure guidance before updating those documents. Done means the guidance and template changes are agreed and complete; the later policy-bot rule is explicitly tracked separately.