docs(ens): R20 Task D — ENSIP-26 follow-up issue + judge evidence (#468)

Opened ensdomains/ensips#72: "ENSIP-26 (Agent Identity Records)
— discussion: 5 design questions for community feedback."

Converts the silent ENSIP-26 PR review into an active discussion
thread by asking 5 specific implementation choices the spec
could go either way on:

  Q1 — bare keys vs agent.* namespace
  Q2 — all-7 SHOULD vs required-core + optional-extensions
  Q3 — fixed Ed25519 vs polymorphic pubkey_<scheme>
  Q4 — JSON array capability + open tags vs canonical registry
  Q5 — reputation_score in ENSIP-26 vs split into sibling ENSIP

Each phrased as "X vs Y, trade-offs A and B, which way" — never
"is this a good idea" (that's the PR's job). Reviewers can
React 👍/👎 or comment a stance per question.

Live verifiable example baked into the issue body: paste-runnable
cast probe for the mainnet apex (all 7 records resolve via
PublicResolver) + viem-style CCIP-Read example for the Sepolia
subname (research-agent.sbo3lagent.eth via deployed
OffchainResolver).

Cross-track citations: the issue explicitly references
Uniswap/universal-router#477 (per-command policy gating uses
ENSIP-26 audit_root) + KeeperHub/cli#57 (workflow envelope
sbo3l_audit_root cites back). Tells reviewers "ENSIP-26 is the
namespace the rest of the agentic-DeFi stack cites back to,"
not an isolated proposal.

Why this matters: Dhaiwat said "you still have time" for late
entrants. A live discussion thread converts "we opened a PR"
(silent) into "we're driving a community conversation" (visible
activity). Higher merge probability + better judge perception.

This PR adds judge evidence: docs/proof/ensip-followup-issue.md
pointing at upstream issue #72.

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: B2JK-Industry