PR #926 follow-ups: sync design_docs governance docs with the *.md narrowing nf-4rb ← Beads

Twinned from https://github.com/cfsh/ninefold/issues/929 by tools/beads/import_github.py's reconcile pass.

Suggestions from ninefold-reviewer's review of #926.

#926 narrowed perimeter.py's GOVERNANCE tuple from design_docs/* (whole tree) to design_docs/*.md, and updated notes/merging.md to match. Three other living docs still assert the old, wider claim — "any diff touching design_docs/ is Tier 2 / always waits for @aedanpope" — which is no longer true for a .py-only diff under design_docs/tools/ (design 008 T3's own case):

1. tools/reviewer/CHARTER.md §D (lines ~199-215): "Tier 2 = ... **anything under design_docs/" and "Any diff touching that tree is Tier 2 at minimum**" and "design_docs/** is denied there too — a design PR always waits for @aedanpope." This is the reviewer's own operating charter, so a future design_docs/tools/**.py-only PR gets misclassified tier=2 with misleading "waits for @aedanpope" prose (informational-only, so it doesn't mis-gate the actual merge, but it's still wrong). 2. design_docs/README.md:65-66: "Every PR touching this folder waits for @aedanpope — including an amendment, and including this index." README.md is explicitly the "kept current" exception to the frozen-round convention, so this one is fair game to fix (unlike 005b/006a/007/008, which are historical records left as written). 3. .claude/commands/design.md:254: "It waits for @aedanpope — always. design_docs/** is inside nflib.perimeter's deny list, so no design PR is ever Tier-1 auto-mergeable."

Not asking for edits to 005b/006a/007/008 or design_docs/002 — those are frozen historical rounds per the "left exactly as written" convention (design_docs/README.md), and none of them is the live spec these three are.

Also minor: #926's body says "(@aedanpope confirmed [the intent] on the issue)" — issue #925's timeline shows no comment from aedanpope, only the automatic "mentioned"/"subscribed" events from the @-mention. Worth citing the real channel next time a PR attributes intent to him, if it wasn't a GitHub comment.