policy-catalogue: add Home Assistant / Open Home Foundation entry

[sjh9714]

Summary

Adds policy-catalogue/home-assistant.md for Home Assistant / the Open Home Foundation, closing #15. The policy is very new — added to home-assistant/core on 2026-07-21 in home-assistant/core#176917 — and repo copy pinned to 702a9cb.

Three things in it that look like schema signal rather than another row:

1. It is the only catalogued policy that discloses the project's own AI use. There's a whole "Our use of AI" section: OHF projects run AI review and triage tooling, its comments may be wrong, and here's how to treat them. Every other entry regulates disclosure flowing one way, toward the maintainers. Reciprocal disclosure is a new field candidate, and arguably the one most consistent with this standard's posture.
2. Template bypass is used as an observable proxy for autonomy. "This includes contributions that bypass the provided issue or pull request templates." Whether a contribution was autonomous isn't decidable from the artifact; whether it filled in the template is. It's the same move Godot makes from the other direction, and the only case in the corpus of a prohibition clause substituting a checkable signal for an unanswerable one.
3. The file declares itself non-authoritative — it names an external canonical URL and says the published version wins on conflict. For a standard premised on a machine-readable file at the repo root, that's a live design question: an adopter with an org-level canonical policy needs a way to say "this is a mirror of X" so agents and CI resolve the right text.

Like Bevy, this is an org-level policy inherited by member repos rather than a repo-scoped one. Two of the newest policies in the catalogue are org-scoped.

I left policy-catalogue/README.md untouched deliberately — #21 and #23 also touch that table and three PRs editing the same rows would conflict. Happy to send one follow-up updating the count and rows once you've decided merge order.

AI disclosure (required — we dogfood our own standard)

• AI-assisted — tool and role: drafted by Claude (Claude Code) working under my direction. Every verbatim excerpt was verified programmatically against the pinned source (702a9cb) before submission — 8 of 8 string checks passed.

Attestations (required)

• understands-change — I can explain every part of this entry and the sources behind it.
• reviewed-output — reviewed before submission. Being precise about the mechanism rather than just ticking: I reviewed a structured walkthrough of the entry plus the verification output confirming each quoted string exists at the pinned commit, rather than reading the rendered diff line-by-line.
• ran-tests — N/A (prose only).

Type of contribution

• Policy-catalogue entry (single file in policy-catalogue/)

Sources

• AI_POLICY.md @ 702a9cb — https://github.com/home-assistant/core/blob/702a9cb7289e535927f5279190bcad6ffc5d3fd0/AI_POLICY.md
• Adoption PR — Add the Open Home Foundation AI Policy home-assistant/core#176917
• Referenced from CONTRIBUTING.md and AGENTS.md at the same commit
• Canonical published version — https://developers.home-assistant.io/docs/ai_policy

Maintainership disclosure: I maintain MergeWarden, a checker in the layer-3 space discussed in #20/#22. This entry documents Home Assistant's policy only and makes no claim about any tool.